JP2019529368A - Bispecific immunomodulatory antibodies linked to costimulatory and checkpoint receptors - Google Patents

Abstract

The present invention relates to bispecific heterodimeric immunomodulating antibodies. [Selection] Figure 1

Description

This application is related to US Provisional Patent Application No. 62 / 381,239 filed Aug. 30, 2016, and US Provisional Patent Application No. 62/456, filed Feb. 7, 2017. No. 033, US Provisional Patent Application No. 62 / 479,723 filed on March 31, 2017, and US Patent Application No. 15 / 623,314 filed on June 14, 2017 The contents of which are expressly fully incorporated by reference in their entirety.

SEQUENCE LISTING This application contains a Sequence Listing that is submitted electronically in ASCII format and is hereby incorporated by reference in its entirety. The above ASCII copy created on August 29, 2017 is 067461_5198-WO_SL. The name is txt, and the size is 26,061,282 kilobytes.

  Tumor-reactive T cells lose their cytotoxic activity over time due to upregulation of suppressive immune checkpoints such as PD-1 and CTLA-4. Two parallel treatment strategies are being pursued so that tumor-reactive T cells can be derepressed and they can continue to kill the tumor cells.

  The first approach is to treat with an antagonistic monoclonal antibody that links to either immunomodulation itself (PD-1, CTLA-4, etc.) or its ligand (PD-L1, PD-L2, CD80, CD86, etc.). Is an immune regulatory block by immunity, which eliminates inhibitory signals that prevent tumor-reactive T cells from killing tumor cells. CTLA-1, PD-1 (programmed cell death 1), TIM-3 (T cell immunoglobulin and mucin domain 3), LAG-3 (lymphocyte activation gene 3), TIGIT (T cells containing Ig and ITIM domains) Immunoregulatory receptors, such as immunoreceptors), and others inhibit the activation, proliferation, and / or effector activity of T cells and other cell types. Preclinical and clinical trials of anti-CTLA4 and anti-PD1 antibodies, including nivolumab, pembrolizumab, ipilimumab, and tremelimumab, led to the hypothesis that immunoregulatory receptors suppress the endogenous T cell response to tumor cells Demonstrated an excellent anti-tumor response, stimulating endogenous T cells to attack tumor cells and providing long-term cancer remission in some patients with various malignancies. Unfortunately, only some patients respond to these therapies, depending on the indication and other factors, but response rates are generally in the range of 10-30%, sometimes monotherapy Every time it gets higher.

  A second approach to derepress tumor-reactive T cells is treated with antagonistic antibodies linked to costimulatory proteins such as ICOS, thereby positively overcoming negative signaling of immune checkpoints. T cell costimulation by adding a signal.

  Accordingly, the present invention relates to costimulatory receptors (eg, ICOS, GITR, OX40, 4-1BB) and checkpoint receptors (eg, PD-1, PD-L1, CTLA-4, LAG-3, TIM-3). Relates to bispecific antibodies that link to BTLA and TIGIT.

Accordingly, in one aspect, the invention provides a dual linking monovalent to human costimulatory receptor and monovalently linking to a human checkpoint receptor for use in activating T cells for the treatment of cancer. Specific antibodies are provided.

  In some embodiments, the costimulatory receptor is selected from the group consisting of ICOS, GITR, OX40, and 4-1BB.

  In a further aspect, the checkpoint receptor is selected from the group consisting of PD-1, PD-L1, CTLA-4, LAG-3, TIGIT, and TIM-3.

  In a further aspect, the antibody is ICOS and PD-1, ICOS and CTLA-4, ICOS and LAG-3, ICOS and TIM-3, ICOS and PD-L1, ICOS and BTLA, ICOS and TIGIT, GITR and TIGIT, GITR. And PD-1, GITR and CTLA-4, GITR and LAG-3, GITR and TIM-3, GITR and PD-L1, GITR and BTLA, OX40 and PD-1, OX40 and TIGIT, OX40 and CTLA-4, IC OX40 OS and LAG-3, OX40 and TIM-3, OX40 and PD-L1, OX40 and BTLA, 4-1BB and PD-1, 4-1BB and CTLA-4, 4-1BB and LAG-3, 4-1BB And TIM-3, 4-1BB and PD-L1, TIGIT and 4-1BB, and 4-1BB and BTLA Ligate to a selected pair of antigens.

  In a further embodiment, the bispecific antibody has a format selected from those outlined in FIG.

  In a further aspect, the invention is a heterodimeric antibody comprising: a) a first heavy chain comprising a first Fc domain, an optional domain linker, and a scFv linked to a first antigen. B) a heavy chain constant domain comprising a second Fc domain, a second heavy chain comprising a heavy chain comprising a hinge domain, a CH1 domain and a variable heavy chain domain; and c) a variable light chain domain and A light chain comprising a light chain constant domain, wherein said variable heavy chain domain and said variable light chain domain form a second antigen linking domain linked to a second antigen, said first and second A heterodimeric antibody is provided wherein one of the antigen-binding domains is linked to human ICOS and the other is linked to human PD-1.

  In a further aspect, the invention provides a heterodimeric bispecific antibody comprising: a) a first heavy chain, i) a first variant Fc domain, and ii) a first antigen. A single chain Fv region (scFv) to be linked, the scFv region comprising a first variable heavy chain, a variable light chain, and a charged scFv linker, wherein the charged scFv linker comprises the first variable heavy chain and the variable light chain. A first heavy chain that covalently links the chain; and b) a second heavy chain comprising a VH-CH1-hinge-CH2-CH3 monomer, wherein VH is a second variable heavy chain A second heavy chain wherein CH2-CH3 is a second mutant Fc domain, and c) a light chain, wherein the second mutant Fc domain is an amino acid substitution N208D / Q295E / N384D / Q418E / N241D, the first and the first Each of the mutant Fc domains comprises amino acid substitutions E233P / L234V / L235A / G236del / S267K, the first mutant Fc domain comprises amino acid substitution S364K / E357Q, and the second mutant Fc domain comprises amino acid substitutions. L368D / K370S, one of the first and second antigen-binding domains linked to human ICOS and the other linked to human PD-1, where the numbering is according to the EU index similar to Kabat, heterodimer Body bispecific antibodies are provided.

  In some embodiments, the heterodimeric antibody has first and second variant Fc domains that each comprise M428L / N434S.

  In a further aspect, the invention is a heterodimeric antibody comprising a) a first heavy chain, i) a first variant Fc domain, and ii) a single chain linked to a first antigen. An Fv region (scFv), wherein the scFv region comprises a first variable heavy chain, a variable light chain, and a charged scFv linker, wherein the charged scFv linker shares the first variable heavy chain and the variable light chain A first heavy chain to be linked; and b) a second heavy chain comprising a VH-CH1-hinge-CH2-CH3 monomer, where VH is a second variable heavy chain, and CH2-CH3 Wherein the first and second mutant Fc domains are L368D / K370S: S364K / E357Q, L368D, wherein the second heavy chain is a second mutant Fc domain and c) a light chain. / K370S: S364K, L368 A set of heterodimerization mutations selected from the group consisting of: / K370S: S364K, T411E / K360E / Q362E: D401K, and T366S / L368A / Y407V: T366W, One is linked to human ICOS and the other is linked to human PD-1, where the numbering provides a heterodimeric antibody according to the EU index similar to Kabat.

  In a further aspect, the present invention provides a nucleic acid composition comprising a nucleic acid encoding a heterodimeric antibody of the present invention, an expression vector composition comprising the nucleic acid, and a host cell comprising the expression vector composition.

  In a further aspect, the present invention provides heterodimeric antibodies for use in the activation of T cells for the treatment of cancer.

About bladder cancer, breast cancer, colon cancer, head and neck cancer, kidney cancer, lung adenocarcinoma, lung squamous cell carcinoma, ovarian cancer, pancreatic cancer, prostate cancer, and melanoma cancer, edited from The Cancer Genome Atlas (TCGA) Presents PD-1 and T cell costimulatory receptor expression data (RNAseq V2 RSEM). The square of the Pearson correlation coefficient was calculated for PD-1 for T cell costimulatory receptors. Figure 2 shows several formats for the bispecific antibodies of the invention. The first is a “bottle opener” format having first and second anti-antigen linking domains. Furthermore, the formats of mAb-Fv, mAb-scFv, center-scFv, center-Fv, center of one arm-scFv, one scFv-mAb, scFv-mAb, dual scFv are all shown. FIG. 2J shows a “center-scFv2” format with two Fab-scFv arms, where the Fab is linked to the first antigen and the scFv is linked to the second antigen. FIG. 2K shows a bispecific mAb format with a first Fab arm linking to the first antigen and a second Fab arm linking to the second antigen. FIG. 2L shows the format of DVD-IgG (see, eg, US Pat. No. 7,612,181, which is expressly incorporated herein by reference and discussed below). FIG. 2M shows a trident format (see, eg, WO2015 / 184203, which is expressly incorporated herein by reference and discussed below). For all the scFv domains shown, they can be either variable heavy chain- (optional linker) -variable light chain, or vice versa, from N-terminus to C-terminus. Furthermore, in a 1-arm scFv-mAb, the scFv can be linked to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain. It is a bottle opener format with ICOS on the Fab side ([ICOS] _H0.66_L0), PD-1 as scFV (1G6_L1.94_H1.279), and including M428L / 434S mutation to prolong serum half-life The sequence of XENP23104 is shown. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers, including uncharged or negatively charged linkers shown in FIG. 8), and slashes indicate variable domain boundaries. Furthermore, the naming convention indicates the orientation of the scFv from the N-terminus to the C-terminus, and some of the sequences herein are oriented as V _H -scFv linker-V _L (from N to the C-terminus), while Are oriented with a _VL- scFv linker- _VH (from N to C-terminus), but as will be appreciated by those skilled in the art, these sequences are also described in their description and Can be used in the opposite orientation. The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs but also the CDRs contained within the _VH and _VL domains using other numbering systems are also included herein. A useful pair of heterodimerized mutation sets (including skew and pI mutations) is shown. In FIGS. 4C and F there are mutations for which there is no corresponding “monomer 2” mutation, which can be used alone for either monomer, or for example the Fab side of a bottle opener For example, a charged scFv linker suitable for a second monomer that utilizes scFv as the second antigen linking domain can be used. A suitable charged linker is shown in FIG. A list of equivalent mutant antibody constant regions and their respective substitutions is shown. pI _ (−) indicates a low pI mutation and pI _ (+) indicates a high pI mutation. These can optionally and independently be combined with other heterodimerization mutations of the invention (and other mutations as outlined herein). Useful removal mutations that have removed FcγR linkage (often referred to as “knockout” or “KO” mutations). In general, deletion mutations are found in both monomers, but in some cases they may be in only one monomer. Two particularly useful embodiments of the invention are shown. Although the “non-Fv” component of this embodiment is shown in FIG. 9A, other formats in FIG. 9 can be used as well. As a component, several charged scFv linkers used in increasing or decreasing the pI of heterodimeric antibodies that utilize one or more scFv are shown. (+ H) positive linkers are specifically used herein. A single prior art scFv linker with a single charge is described in Whitlow et al. , Protein Engineering 6 (8): 989-995 (1993). It should be noted that this linker was used to reduce aggregation in scFv and increase proteolytic stability. The sequences of several useful bottle opener format scaffolds are shown based on human IgG1, without Fv sequences (eg, vh and vl for scFv and Fab sides). Skeletal 1 of the bottle opener is based on human IgG1 (356E / 358M allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del / S267K contains both strand removal mutations. Skeletal 2 of the bottle opener is based on human IgG1 (356E / 358M allotype), with different skew mutations, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and both E233P / L234V / L235A / G236del / S267K Contains strand removal mutations. Bottle Opener Skeleton 3 is based on human IgG1 (356E / 358M allotype), with different skew mutations, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and both E233P / L234V / L235A / G236del / S267K Contains strand removal mutations. The bottle opener scaffold 4 is based on human IgG1 (356E / 358M allotype) and has different skew mutations, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and both E233P / L234V / L235A / G236del / S267K Contains strand removal mutations. Skeletal 5 of the bottle opener is based on human IgG1 (356D / 358L allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del / S267K contains both strand removal mutations. The skeleton 6 of the bottle opener is based on human IgG1 (356E / 358M allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del / S267K both strand removal mutations, as well as N297A both strand mutations. The skeleton 7 of the bottle opener is identical to 6 except that the mutation is N297S. Alternative formats for the bottle opener scaffolds 6 and 7 can exclude both strand removal mutations E233P / L234V / L235A / G236del / S267K. Skeletal 8 is based on human IgG4, S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and E233P / L234V / L235A / G236del / S267K both chains It includes a deletion mutation as well as a mutation in both strands of S228P (EU numbering, which is S241P in Kabat) that eliminates Fab arm exchange, as is known in the art. In an alternative format for the skeleton 8 of the bottle opener, both E233P / L234V / L235A / G236del / S267K strand removal mutations can be excluded, and skeleton 9 is based on human IgG2 and S364K / E357Q: skew of L368D / K370S. Mutation, including a pI mutation on the Fab side of N208D / Q295E / N384D / Q418E / N421D. Skeletal 10 is based on human IgG2 and contains S364K / E357Q: S368D / K370S skew mutation, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and S267K both chain mutations. As understood by those skilled in the art and outlined below, these sequences include one monomer containing scFv (optionally including a charged scFv linker) and another monomer containing a Fab sequence, It can be used with any of the vh and vl pairs outlined herein (eg, vh linked to a “Fab heavy chain” and vl linked to a “constant light chain”). That is, any of the Fv sequences outlined herein for anti-CTLA-4, anti-PD-1, anti-LAG-3, anti-TIM-3, anti-TIGIT, and anti-BTLA are scFv (also optionally charged These can be incorporated in any combination into the backbone of FIG. 37, whether or not as having an scFv linker) or as a Fab. The constant light chain shown in FIG. 9A can be used for all of the constructs in the figure, but the kappa constant light chain may also be substituted. Note that these bottle opener scaffolds are used in the center-scFv format of FIG. 1F, where an additional second Fab (vh-CH1) having the same antigenic linkage as the first Fab. And vl-constant light chain) are added to the N-terminus of the “bottle opener side” scFv. Included in each of these backbones is 90, 95, 98, and 99% identical to the listed sequence (as defined herein) and / or (to those skilled in the art As can be seen, 1, 2, 3, 4, 5, compared to the “parent” in the figure which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4 based on the backbone) A sequence that includes 6, 7, 8, 9, or 10 additional amino acid substitutions, ie, the listed backbone contains additional amino acid modifications in addition to the skew, pI, and excision mutations contained within the backbone of this figure (Generally amino acid substitutions). The sequence of a selected number of anti-PD-1 antibodies is shown. It is important to note that these sequences were generated based on human IgG1 with the excision mutation shown in FIG. 6A (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”) . CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein. A selected number of PD-1 ABDs are shown, along with additional anti-PD-1 ABDs listed as SEQ ID NOs: 1-239, 3125-3144, 4697-7594, and 4697-21810. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers, including uncharged or negatively charged linkers shown in FIG. 8), and slashes indicate variable domain boundaries. Furthermore, the naming convention shows the orientation of scFv from N-terminus to C-terminus, some of the sequences in this figure are oriented as VH-scFv linker-VL (from N to C-terminus), while Are oriented with the VL-scFv linker-VH (from N to C-terminus), but as will be appreciated by those skilled in the art, these sequences are also oriented in the opposite direction to their depiction herein. Can be used in The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format. A selected number of CTLA-4 ABD is shown, with additional anti-CTLA-4 ABD listed as SEQ ID NOs: 2393-2414 and 3737-3816. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers, including uncharged or negatively charged linkers shown in FIG. 8), and slashes indicate variable domain boundaries. Furthermore, the naming convention shows the orientation of scFv from N-terminus to C-terminus, some of the sequences in this figure are oriented as VH-scFv linker-VL (from N to C-terminus), while Are oriented with the VL-scFv linker-VH (from N to C-terminus), but as will be appreciated by those skilled in the art, these sequences are also oriented in the opposite direction to their depiction herein. Can be used in The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format. A selected number of LAG-3 ABD is shown, with additional anti-LAG-3 ABD listed as SEQ ID NOs: 2415-2604 and 3817-3960. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers including uncharged or negatively charged linkers shown in FIG. 8), and the slash indicates the variable domain> boundary (s). Furthermore, the naming convention shows the orientation of scFv from N-terminus to C-terminus, some of the sequences in this figure are oriented as VH-scFv linker-VL (from N to C-terminus), while Are oriented with the VL-scFv linker-VH (from N to C-terminus), but as will be appreciated by those skilled in the art, these sequences are also oriented in the opposite direction to their depiction herein. Can be used in The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format. The sequence of a selected number of anti-TIM-3 antibodies is shown. Note that these sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs but also the CDRs contained within the _VH and _VL domains using other numbering systems are also included herein. The sequence of a selected number of anti-PD-L1 antibodies is shown. These sequences were generated based on the human IgG1 backbone with the deletion mutations outlined herein (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats as well It is important to note that can be used. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein. The sequence of the prototype anti-4-1BB antibody is shown. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein. The sequence of the prototype anti-OX40 antibody is shown. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein. The sequence of the prototype anti-GITR antibody is shown. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein. The sequence of the prototype anti-ICOS antibody is shown. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR positions may vary slightly depending on the numbering used as shown in Table X, as noted herein and for any sequence containing CDRs herein, so Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein. The sequence of an exemplary anti-ICOS Fab is shown. CDRs are underlined and slashes (/) indicate variable region boundaries. The exact identification of the CDR positions may vary slightly depending on the numbering used as shown in Table X and is underlined, as applies to any sequence described herein and including the CDRs herein. CDRs included in VH and VL domains using other numbering systems as well as other CDRs are also included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format. It is important to note that these sequences were generated using a 6-histidine (His6 or HHHHHH) C-terminal tag at the C-terminus of the heavy chain being removed. Figure 7 shows the change in melting temperature from the parent Fab (XENP22050) as determined by DSF of a mutant anti-ICOS Fab modified for melting temperature ( _Tm ) and stability. The equilibrium dissociation constant (KD), association rate (Ka), and dissociation rate (Kd) of the mutant anti-ICOS Fab for the mouse Fc fusion of human ICOS captured on the AMC biosensor as determined by Octet Show. Equilibrium dissociation constant (KD), association rate (KA), and dissociation rate of mutant anti-ICOS Fab for biotinylated IgG1 Fc fusion of human ICOS captured on SA biosensor as determined by Octet ( Kd). 2 shows an exemplary anti-ICOS scFv sequence. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) 4 linker, but as will be appreciated by those skilled in the art, this linker is , Some of which may be replaced by other linkers including uncharged or negatively charged linkers shown in Figure X), with a slash (/) indicating the boundary of the variable region. The nomenclature indicates the orientation of scFv from N-terminus to C-terminus, some of the sequences in this figure are oriented with VH-scFv linker-VL (from N to C-terminus, see FIG. 24) While some are oriented with the VL-scFv linker-VH (N to C-terminus, see FIG. 24B), as will be appreciated by those skilled in the art, these sequences are also described herein. It can be used in an orientation opposite to their depiction. The exact identification of the CDR positions may vary slightly depending on the numbering used as shown in Table X, as noted herein and for any sequence containing CDRs herein, so Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format. It is important to note that these sequences were generated using a polyhistidine (His6 or HHHHHH) C-terminal tag at the C-terminus of the heavy chain being removed. Mutant anti-ICOS scFv modified for melting temperature (Tm) and stability, parent scFv as determined by DSF (XENP24352, VH-scFv linker-VL orientation from N-terminus to C-terminus) Shows the change in melting temperature from. Figure 2 shows the amino acid sequence of a prototype anti-costim x anti-checkpoint antibody in bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain is either a different orientation of V _H -scFv linker-V _L or V _L -scFv linker-V _H (from N-terminus to C-terminus) but can reverse this it can. Furthermore, each sequence outlined herein can include or exclude a mutation of M428L / N434S in one or preferably both Fc domains, which results in a longer half-life in serum. Induction of cytokine secretion by a prototype costim / checkpoint bottle opener in the SEB stimulated PBMC assay. FIG. 5 shows SEB-stimulated PBMC after treatment with a test sample indicated as induction of IL-2 secretion in naive (non-SEB stimulation). Since the tumor TIL co-expresses immune checkpoint receptors and costimulatory receptors, bispecific antibodies are shown to increase specificity, enhance antitumor activity, and avoid peripheral toxicity. Schematics related to the benefits of costim × checkpoint blocking bispecific antibodies are shown. Double positive cells are selectively occupied by an exemplary anti-ICOS × anti-PD-1 antibody (XENP20896) compared to a one-arm anti-PD-1 antibody (XENP20111) and a one-arm anti-ICOS antibody (XENP20266). It has been. Anti-ICOS × anti-PD-1 bispecific antibody against A) PD-1 and ICOS double positive T cells and B) PD-1 and ICOS double negative T cells after SEB stimulation of human PBMC (XENP20896) shows the receptor occupancy of 1-arm anti-ICOS antibody (XENP20266) and 1-arm anti-PD-1 antibody (XENP20111). 2 shows the amino acid sequence of an exemplary anti-ICOS × anti-PD-1 antibody in a bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has a different orientation of either V _H -scFv linker-V _L or V _L -scFv linker-V _H (from N-terminus to C-terminus) but reverse this be able to. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. The amino acid sequence of an exemplary anti-ICOS × anti-PD-1 antibody in a bottle opener format containing a mutation of M428L / N434S that results in a longer half-life in serum in one or preferably both Fc domains is shown. Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has a different orientation of either V _H -scFv linker-V _L or V _L -scFv linker-V _H (from N-terminus to C-terminus) but reverse this be able to. Equilibrium dissociation constant (KD), association rate (Ka) of mutant anti-ICOS × anti-PD-1 bispecific antibody for mouse Fc fusion of human ICOS captured on AMC biosensor as determined by Octet ) And dissociation rate (Kd). Equilibrium dissociation constant (KD) of mutant anti-ICOS × anti-PD-1 bispecific antibody for biotinylated IgG1 Fc fusion of human and ICOS captured on SA / SAX biosensor as determined by Octet ), Association rate (KA), and dissociation rate (Kd). FIG. 5 shows cell surface ligation of mutant anti-ICOS × anti-PD-1 bispecific antibody to human T cells in SEB stimulated PBMC assay in two separate experiments shown in A) and B). Mutant anti-ICOS × anti-PD-1 bispecific antibody, 1-arm anti-ICOS antibody, and 1-arm anti-PD-1 antibody against PD-1 and ICOS double-positive T cells after SEB stimulation of human PBMC The receptor occupancy of (XENP20111) is shown. FIG. 6 shows that the mutant anti-ICOS × anti-PD-1 bispecific antibody promotes secretion of A) IL-2 and B) IFNγ secreted from SEB-stimulated PBMC. FIG. 6 shows that the mutant anti-ICOS × anti-PD-1 bispecific antibody promotes secretion of A) IL-2 and B) IFNγ secreted from SEB-stimulated PBMC. The concentration of IFNγ at A) day 7 and B) day 11 after transplantation of human PBMC and treatment with the indicated test samples is shown. Shown are CD45 + cell numbers in mice as determined by flow cytometry at A) 11 days and B) 14 days after transplantation of human PBMC and treatment with the indicated test samples. Shown are the numbers of A) CD8 + T cells and B) CD4 + T cells in mice as determined by flow cytometry on day 14 after transplantation of human PBMC and treatment with the indicated test samples (** P ≦ 0. 01). Shows changes in mouse body weight up to 14 days after transplantation of human PBMC and treatment with the indicated test samples (** P ≦ 0.01). Shows the concentration of IFNγ at A) 7 days and B) 14 days after transplantation of human PBMC and treatment with the indicated test samples Shown is the number of CD45 + cells in mice as determined by flow cytometry on day 14 after transplantation of human PBMC and treatment with the indicated test samples. Shown are the numbers of A) CD8 + T cells and B) CD4 + T cells in mice as determined by flow cytometry on day 14 after transplantation of human PBMC and treatment with the indicated test samples. Shows changes in mouse body weight up to day 12 and day 15 after transplantation of human PBMC and treatment with the indicated test samples. FIG. 2 shows the amino acid sequence of an exemplary anti-ICOS × anti-PD-1 antibody and alternative ICOS ABD in bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. FIG. 6 shows IL-2 cytokine release assay after SEB stimulation of human PBMC and treatment with alternative anti-ICOS × anti-PD-1 bispecific antibody. FIG. 5 shows IL-2 cytokine release assay (as fold induction over bivalent anti-RSV mAb) following SEB stimulation of human PBMC and treatment with alternative anti-ICOS × anti-PD-1 bispecific antibody. AKT phosphorylation in SEB stimulated purified CD3 + T cells after treatment with anti-ICOS × anti-PD-1 bispecific antibody. A) IL-17A, B) IL17F, C) IL-22, D) IL-10, E) IL with the indicated test samples upon induction with the indicated bivalent anti-RSV as determined by NanoString -9, and F) Fold induction of IFNγ gene expression is shown. Shown is the fold of mean induction of expression of selected immune response genes by the indicated test sample upon treatment with bivalent anti-RSV mAb as determined by NanoString. The shading intensity corresponds to the magnitude of the magnification change. Shown is the number of CD45 + cells in mice as determined by flow cytometry on day 14 after transplantation of human PBMC and treatment with the indicated test samples. Similar to FIGS. 9 and 75, several, including the central scFv of FIG. 2F, the central-scFv2 format of FIG. 2J, the bispecific mAb of FIG. 2K, the DVD-Ig of FIG. 2L, and the trident format of FIG. The sequence of the “skeleton” position (eg, not including Fv) in the additional format is shown. In FIG. 2L, the DVD-Ig® linker is shown double underlined with other linkers found in WO2007 / 024715, the entirety of which is specifically incorporated herein by reference, particularly for those sequences. Embedded in the book. In the trident format, other trident linkers and coil-coil arrangements are shown in WO2015 / 184203, which is hereby incorporated by reference in its entirety, particularly for those sequences. As will be appreciated by those skilled in the art, bold domains (eg, “VH1”, VH2-scFv linker-VL2 ”, etc.) are separated by a slash“ / ”and may optionally contain any domain linker. All of these scaffolds utilize a kappa constant region in the light chain, although lambda chains can also be used. As outlined herein, with respect to FIGS. 9 and 75, these scaffolds can be combined with any VH and VL domains. 2 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. 2 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in a center-scFv format. Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. 2 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in a center-scFv2 format. Antibodies are named using the Fab variable region first, scFv variable region second, followed by the chain designation (heavy or light chain). CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. 2 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in a bispecific mAb format. The antibody comprises a first Fab variable region for a first antigen, a second Fab variable region for a second antigen, separated by a dash, followed by a chain designation (heavy chain 1 or 1 for the first antigen). Named using light chain 1, heavy chain 2 or light chain 2) for second antigen. CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. 2 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in DVD-IgG format. The antibodies are named using a first variable region for a first antigen, a second Fab variable region for a second antigen, followed by a chain designation (heavy or light chain). CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. FIG. 2 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in a trident format. Antibodies are separated by dashes using the first VL and VH for the first antigen, including DART, the Fab variable region for the second antigen, followed by the chain designation (heavy or light chain). Named. CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. In the SEB-stimulated PBMC assay, the induction of cytokine secretion (IL-2) by an alternative format costim x checkpoint blocking bispecific antibody is shown. 2 shows the amino acid sequence of an exemplary anti-ICOS × anti-CTLA-4 antibody in a bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. 2 shows the amino acid sequence of an exemplary anti-LAG-3 × anti-ICOS antibody in a bispecific mAb format. The antibody comprises a first Fab variable region for a first antigen, a second Fab variable region for a second antigen, separated by a dash, followed by a chain designation (heavy chain 1 or 1 for the first antigen). Named using light chain 1, heavy chain 2 or light chain 2) for second antigen. CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. 2 shows the amino acid sequence of an exemplary anti-TIM-3 × anti-ICOS antibody in a bispecific mAb format. The antibody comprises a first Fab variable region for a first antigen, a second Fab variable region for a second antigen, separated by a dash, followed by a chain designation (heavy chain 1 or 1 for the first antigen). Named using light chain 1, heavy chain 2 or light chain 2) for second antigen. CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. Amino acid sequences of anti-ICOS × anti-PD-L1 antibody in bottle opener format (FAB-scFv-Fc) and central-scFv2 format are shown. The bottle openers are named using the Fab variable region first and scFv variable region second, separated by a dash, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). The Center-scFv2 is named using Fab variable region first, scFv variable region second, followed by chain designation (heavy chain or light chain). CDRs are underlined and slashes indicate variable region boundaries. CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains. FIG. 5 shows the induction of cytokine secretion (IL-2) by additional costim × checkpoint blocking bispecific antibody in the SEB stimulated PBMC assay. 1 shows the amino acid sequence of an exemplary 1-arm anti-ICOS Fab-Fc antibody. CDRs are underlined and slashes indicate variable region boundaries. Since one amino acid chain is only the Fc domain and the other side is the anti-ICOS Fab side, these are referred to as “one arm” or “one arm” format. The Fc domain contains the S364K / E357Q skew mutation, as well as the pI (−)-equivalent_A mutation shown in FIG. The Fab Fc domain contains the L368D / K370S skew mutation and the pI ISO (+ RR) mutation shown in Figure X. Both Fc domains contain a deletion mutation (E233P / L234V / L235A / G236del / S267K). Equilibrium dissociation constant (KD), association rate (Ka), and dissociation of mutant 1-arm anti-ICOS Fab-Fc antibody against mouse Fc fusion of human ICOS captured on AMC biosensor as determined by Octet Speed (Kd) is shown. AKT phosphorylation in SEB-stimulated purified CD3 + T cells after treatment with bivalent and monovalent anti-PD-1 antibody and anti-ICOS × anti-PD-1 bispecific antibody. Figure 3 shows AKT phosphorylation in purified CD3 + T cells after treatment with monovalent anti-ICOS Fab-Fc antibody with alternative anti-ICOS ABD. Prototype bispecific antibodies (OX40 × PD-1, GITR × PD-1, 4-1BB × PD-1, CTLA-4 × ICOS) are shown. Shows some prototype mAbs (4-1BB, OX40, GITR, ICOS, PD-L1, and PD-1), whose Fv is combined with other Fv of the present invention and in any format (bottle opener, mAb-Fv, mAb-scFv, center-scFv, bispecific mAb, center-Fv, one-arm center-scFv, one-arm scFv-mAb, dual scFv, DVD-Ig, or trident) it can. Similarly, some additional ICOS × PD-L1 bottle opener arrays are also shown. Further PD-1 × ICOS bottle openers are shown, in some cases PD-1 Fv is in Fab format, ICOS Fv is in scFv format, and in other cases it is reversed. Figure 2 shows the sequence of the mAb-scFv backbone of use in the present invention, to which the Fv sequence of the present invention has been added. Skeletal 1 of mAb-scFv is based on human IgG1 (356E / 358M allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / It contains a deletion mutation in both strands of L235A / G236del / S267K. Skeletal 2 is based on human IgG1 (356D / 358L allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del Includes removal mutations in both strands of S267K. Skeletal 3 is based on human IgG1 (356E / 358M allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del Includes both strand removal mutations in S267K, as well as mutations in both strands of N297A. Skeletal 4 is identical to 3 except that the mutation is N297S. Alternative formats for mAb-scFv scaffolds 3 and 4 can exclude both strand removal mutations E233P / L234V / L235A / G236del / S267K. Skeletal 5 is based on human IgG4, S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and E233P / L234V / L235A / G236del / S267K both chains Including a mutation in both strands of S228P (EU numbering, which is S241P in Kabat) that eliminates Fab arm exchange, as known in the art, and backbone 6 is based on human IgG2 In addition, the S364K / E357Q: L368D / K370S skew mutation and the N208D / Q295E / N384D / Q418E / N421D pI mutation are included on the Fab side. Skeletal 7 is based on human IgG2 and contains S364K / E357Q: S368D / K370S skew mutation, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and S267K both chain mutations. As understood by those of skill in the art and outlined below, these sequences are composed of one monomer containing both Fab and scFv (optionally including a charged scFv linker) and another monomer containing the Fab sequence. Can be used with any of the vh and vl pairs outlined herein (eg, vh linked to a “Fab heavy chain” and vl linked to a “constant light chain”). That is, any of those outlined herein for anti-CTLA-4, anti-PD-1, anti-LAG-3, anti-TIM-3, anti-TIGIT, anti-BTLA, anti-ICOS, anti-GITR, country OX40, and anti-4-1BB The Fv sequences can be incorporated in any combination into this FIG. 75 backbone, whether or not as scFv (also with a charged scFv linker optionally) or Fab. Monomer 1 side is the pI negative side of Fab-scFv, heterodimerization mutation L368D / K370S, equivalent pI mutation N208D / Q295E / N384D / Q418E / N421D, removal mutation E233P / L234V / L235A / G236del / S267K (All compared to IgG1). The monomer 2 side is the pI positive side of scFv and contains the heterodimerization mutation 364K / E357Q. However, other skew mutation pairs, in particular, [S364K / E357Q: L368D / K370S], [L368D / K370S: S364K], [L368E / K370S: S364K], [T411T / E360E / Q362E: D401K], [L368D / K370S: S364K / E357L], [K370S: S364K / E357Q], [T366S / L368A / Y407V: T366W], and [T366S / L368A / Y407V / Y394C: T366W / S354C]. The constant light chain shown in FIG. 75A can be used for all of the constructs in the figure, but the kappa constant light chain may also be substituted. These mAb-scFv scaffolds are similar in both the scFv-mAb format of FIG. 1H in the mAb-Fv format (where one monomer contains vl at the C-terminus and the other contains vh at the C-terminus). It should be noted that it is used in one of the monomers with the addition of a scFv domain at the C-terminus. Included in each of these backbones is 90, 95, 98, and 99% identical to the listed sequence (as defined herein) and / or (to those skilled in the art As can be seen, 1, 2, 3, 4, 5, compared to the “parent” in the figure which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4 based on the backbone) A sequence that includes 6, 7, 8, 9, or 10 additional amino acid substitutions, ie, the listed backbone contains additional amino acid modifications in addition to the skew, pI, and excision mutations contained within the backbone of this figure (Generally amino acid substitutions). Some prior art sequences of Fv linked to human PD-1 are shown as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD. Some prior art sequences of Fv linked to human ICOS are shown as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fv can be associated with checkpoint receptors (eg, PD-1, PD-L1, CTLA-4, TIM-3, LAG-3, TIGIT, and BTLA). Can be combined with linking Fvs (including Fv sequences included herein) and in any format (bottle opener, mAb-Fv, mAb-scFv, center-scFv, bispecific mAb, center-Fv 1 arm center-scFv, 1 arm scFv-mAb, dual scFv, DVD-Ig, or trident). In particular, they can be combined with PD-1 ABD having identifiers 1G6_H1.279_L1.194, 1G6_H1.280_L1.224, 1G6_L1.194_H1.279, 1G6_L1.210_H1.288, and 2E9_H1L1. As the VH and VL sequences, several prior art sequences of Fv linked to human PD-L1 are shown. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD. Several prior art sequences of Fv linked to human CTLA-4 are shown as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD. Some prior art sequences of Fv linked to human LAG-3 are shown as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD. Some prior art sequences of Fv linked to human TIM-3 are shown as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD. Some prior art sequences of Fv linked to human BTLA are shown as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD. Several prior art sequences of Fv linked to human TIGIT are shown as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD. Several BTLA ABDs are shown with additional anti-BTLA ABDs listed as SEQ ID NOs: 3705-3736. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers, including uncharged or negatively charged linkers shown in FIG. 8), and slashes indicate variable domain boundaries. As noted above, the naming convention indicates the orientation of scFv from N-terminus to C-terminus, and in the sequences listed in this figure they are all vh-scFv linker-vl (from N-terminus to C-terminus). However, these sequences can also be used in the opposite orientation (from N-terminus to C-terminus), vl-linker-vh. The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs but also the CDRs contained within the vh and vl domains using other numbering systems are also included herein. Furthermore, for all sequences in the figure, these vh and vl sequences can be used in either scFv format or Fab format. A matrix of possible combinations of costim and checkpoint ABD, with all possible combinations. “A” in the box means that PD-1 ABD is 1G6_L1.194_H1.279. “B” in the box indicates that ICOS ABD is [ICOS] H. 066_L0. “C” in the box means that PD-1 is the paired scFv. “D” in the box means that CTLA-4 ABD is Fab and is [CTLA-4] _H3_L0.22. “E” in the box means that CTLA-4 ABD is scFv and is [CTLA-4] _H3.23_L0.22. “F” in the box means that LAG-3 ABD is 7G8_H3.30_L1.34. “G” in the box means that BTLA ABD is 9C6_H1.1_L1. “H” in the box means that the combination is a bottle opener. “I” in the box means that the combination is in the center-scFv format. Two additional ICOS × PD-1 bottle openers are shown.

I. Nomenclature The bispecific antibodies of the invention are listed in several different formats. As understood in the art, each polypeptide is given a unique “XENP” number, although longer sequences may include shorter ones. For example, the scFv monomer heavy chain in the bottle opener format for a given sequence may have a first XENP number, but the scFv domain may have a different XENP number. Some molecules have three polypeptides, so the XENP number with the component is used as a name. Thus, the bottle opener format molecule XENP generally includes three sequences or equivalents, referred to as “XENP23104-HC-Fab”, “XENP23104HC-scFv” and “XENP23104LC”, although those of ordinary skill in the art through sequence alignment Can be easily identified. These XENP numbers are in the sequence listing as well as identifiers and are used in the figure. In addition, one molecule containing three components gives rise to multiple sequence identifiers. For example, the list of Fab monomers has three CDRs, a full length sequence, a variable heavy chain sequence, and a variable heavy chain sequence, and a light chain has three CDRs, a full length sequence, a variable light chain sequence, and a variable light chain sequence. The scFv-Fc domain has a full-length sequence, an scFv sequence, a variable light chain sequence, three light chain CDRs, an scFv linker, a variable heavy chain sequence, and three heavy chain CDRs, and has a scFv domain. Note that all molecules in the specification use a single charged scFv linker (+ H), but others can be used. Furthermore, the nomenclature for a particular variable domain uses a format of the “Hx.xx_Ly.yy” type, where the number is a unique identifier for a particular variable chain sequence. Therefore, the variable domain on the scFv side of XENP23104 (linked to PD-1) is “1G6_L1.194_H1.279”, which means that the variable heavy chain domain H1.279 was combined with the light chain domain L1.194. Indicates. When these sequences are used as scFv, the designation “1G6_L1.194_H1.279” indicates that the variable heavy chain domain H1.279 is combined with the light chain domain L1.194, from the N-terminus toward the C-terminus. Indicates vl-linker-vh orientation. This molecule having the same sequence of heavy and light chain variable domains but in reverse order will be named “1G6_H1.279_L1.194”. Similarly, different constructs can “mix and match” heavy and light chains, as will be apparent from the sequence listing and figures.

II. Incorporation of materials A. Figures and illustrations Specifically incorporated by reference are the figures, illustrations and arrangements from USSN 62 / 479,723, and the illustrations, illustrations and arrangements from USSN 15 / 623,314. Furthermore, the claims from USSN 62 / 479,723 are more specifically incorporated.

B. Sequence Target antigen: The sequence of human PD-1 (sp | Q15116) is SEQ ID NO: 26226. The sequence of the extracellular domain of human PD-1 (sp | Q15116 | 21-170) is SEQ ID NO: 26227. The sequence of Macaca fascicularis PD-1 (tr | B0LAJ3) is SEQ ID NO: 26228. The sequence (predicted) (tr | B0LAJ3 | 21-170) of the extracellular domain of PD-1 of Macaca fascicularis is SEQ ID NO: 26229. The sequence of human CTLA-4 (sp | P16410) is SEQ ID NO: 26230. The sequence of the extracellular domain of human CTLA-4 (sp | P16410 | 36-161) is SEQ ID NO: 26231. The sequence (tr | G7PL88) of CTLA-4 of Macaca fascicularis is SEQ ID NO: 26232. The sequence (predicted) (tr | G7PL88) of the extracellular domain of Macaca fascicularis CTLA-4 is SEQ ID NO: 26233. The sequence of human LAG-3 (sp | P18627) is SEQ ID NO: 26234. The sequence of the extracellular domain of human LAG-3 (sp | P18627 | 29-450) is SEQ ID NO: 26235. The sequence (predicted) of LAG-3 of Macaca fascicularis (gi | 544447815 | ref | XP_005570011.1) is SEQ ID NO: 26236. The sequence (predicted) of the extracellular domain of Macaca fascicularis LAG-3 (gi | 544447785 | ref | XP_005570011.1 | 29-450) is SEQ ID NO: 26237. The sequence of human TIM-3 (sp | Q8TDQ0) is SEQ ID NO: 26238. The sequence of the extracellular domain of human TIM-3 (sp | Q8TDQ0 | 22-202) is SEQ ID NO: 26239. The sequence (predicted) of TIM-3 of Macaca fascicularis (gi | 355570365 | gb | EHH54703.1) is SEQ ID NO: 26240. The sequence (predicted) of the extracellular domain of TIM-3 of Macaca fascicularis (gi | 355570365 | gb | EHH54703.1 | 22-203) is SEQ ID NO: 26241. The sequence of human PD-L1 (sp | Q9NZQ7) is SEQ ID NO: 26242. The sequence of the extracellular domain of human PD-L1 (sp | Q9NZQ7 | 19-238) is SEQ ID NO: 26243. The sequence (prediction) of PD-L1 of Macaca fascicularis (gb | XP_005583836.1) is SEQ ID NO: 26244. The sequence (prediction) (gb | XP_005583836.1 | 19-238) of the extracellular domain of Macaca fascicularis PD-L1 is SEQ ID NO: 26245. The sequence of human ICOS (sp | Q9Y6W8) is SEQ ID NO: 26246. The sequence of the extracellular domain of human ICOS (sp | Q9Y6W8 | 21-140) is SEQ ID NO: 26247. The sequence of macOS fascicularis ICOS (gi | 5444477053 | ref | XP_005574075.1) is SEQ ID NO: 26248. The sequence (predicted) of the extracellular domain of MacOS fascicularis ICOS (gi | 5444477053 | ref | XP_005574075.1 | 21-140) is SEQ ID NO: 26249. The sequence of human GITR (sp | Q9Y5U5) is SEQ ID NO: 26250. The sequence of the extracellular domain of human GITR (sp | Q9Y5U5 | 26-162) is SEQ ID NO: 26251. The sequence (prediction) of the GITR of Macaca fascicularis (ref | XP_005545180.1) is SEQ ID NO: 26252. The sequence (predicted) of the extracellular domain of GITR of Macaca fascicularis (ref | XP_005545180.1 | 26-162) is SEQ ID NO: 26253. The sequence of human OX40 (sp | P43489) is SEQ ID NO: 26254. The sequence of the extracellular domain of human OX40 (sp | P43489 | 29-214) is SEQ ID NO: 26255. The sequence (prediction) of OX40 of Macaca fascicularis (ref | XP_005545179.1) is SEQ ID NO: 26256. The sequence (predicted) of the extracellular domain of Macaca fascicularis OX40 (ref | XP_005545179.1 | 29-214) is SEQ ID NO: 26257. The sequence of human 4-1BB (sp | Q07011) is SEQ ID NO: 26258. The sequence of the extracellular domain of human 4-1BB (sp | Q07011 | 24-186) is SEQ ID NO: 26259. The sequence (prediction) of 4-1BB of Macaca fascicularis (ref | XP_005544945.1) is SEQ ID NO: 26260. The sequence (predicted) of the extracellular domain of Macaca fascicularis 4-1BB (ref | XP_00554495.45.1 | 24-186) is SEQ ID NO: 26261.

ICOS Linking Domain: Further, the sequences shown in FIGS. 19, 20, and 24, SEQ ID NOs: 27869-28086, as indicated by the nomenclature, are several ICOS Fab sequences (heavy chain VH1-CH1 and light chain). VL1-CL). References to junctions between CDRs and variable junctions are shown in FIG. 40 of USSN 62 / 479,723 (incorporated herein by reference, the illustration is similar), but those skilled in the art will also see from the sequence listing. CDRs (see Table 1 through numbering and / or sequence alignment), and junctions (eg, heavy chain CH1 generally starts with the sequence “ASTK ...” and light chain constant domains start with “RTVA ...”; As shown in the nomenclature, SEQ ID NOs: 28087-28269 are three sequences of “one arm mAb” (FIG. 2N; Fab-Fc, Fc only, and light chain). References to junctions between CDRs and variable junctions are shown in FIG. 41 of USSN 62 / 479,723 (incorporated herein by reference). However, those skilled in the art will also recognize CDRs (see Table 1 through numbering and / or sequence alignment) from the sequence listing, and junctions (eg, heavy chain CH1 is generally in the “ASTK ...” sequence). It can be seen that the light chain constant domain starts with “RTVA ...” An additional one-arm ICOS molecule is shown in Figure 68. SEQ ID NOs: 28549-28556 are similar Shows some control antibodies (HC and LC) that can also use Fv as an ICOS ABD, and references to conjugation between CDRs and variable junctions are shown in FIG. 44 of USSN 62 / 479,723 (Incorporated herein by reference, the illustration is similar), but those skilled in the art will also recognize CDRs (numbering and / or sequencing) from the sequence listing. Through the alignment (see Table 1), and the junction (eg, heavy chain CH1 generally starts with the sequence “ASTK ...” and light chain constant domains start with “RTVA ...”). SEQ ID NOs: 28557-28665 show several ICOS scFvs used as combinations in the present invention, and references to junctions between CDRs and variable junctions are shown in FIG. 45 of USSN 62 / 479,723 (Incorporated herein by reference, the illustration is similar), but one of skill in the art can also identify CDRs (see Table 1 through numbering and / or sequence alignment) from the sequence listing, and scFv is vh And a charged linker (GKPGS) ₄ is used between It will be. Thus, ICOS ABDs suitable for use in combination with ABD for the checkpoint receptor are FIGS. 19, 20, 24, 68, and 77 and SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665, and [ICOS] _H0.66_L0 and [ICOS] _H0_L0.

  In addition, suitable ICOS × PD-1 bottle opener sequences are those in FIG. 3 and FIGS. 42 and 43 of USSN 62 / 479,723, both of which are incorporated herein by reference, as well as their illustrations and Three figures are used to show CDRs, junctions, etc. as well, including those in SEQ ID NOs: 28270-28548 corresponding to the sequences shown in the sequence also incorporated).

III. Summary Therapeutic antibodies against immunomodulatory inhibitors of immunity, such as PD-1, are very promising in limited situations in the clinic for cancer treatment. Cancer can be considered that the patient has no ability to recognize and eliminate cancerous cells. In many cases, these transformed (eg, cancerous) cells counteract immune surveillance. There are natural regulatory mechanisms that limit T cell activation in the body to prevent unrestricted T cell activity, which can be utilized by cancerous cells to avoid or suppress immune responses. It is the purpose of immunotherapy to restore the ability of immune effector cells, particularly T cells, to recognize and eliminate cancer. The field of immuno-oncology, sometimes referred to as “immunotherapy”, is evolving rapidly, with T cell checkpoint-inhibiting antibodies such as Yervoy, Keytruda, and Opdivo having several recently approved ones. is there. These antibodies are commonly referred to as “checkpoint inhibitors” because they block the normally negative regulators of T cell immunity. It is generally understood that a variety of immunoregulatory signals, both costimulatory and costimulatory, can be used to integrate optimal antigen-specific immune responses.

  A checkpoint inhibitor monoclonal antibody is linked to an immunomodulatory inhibitor protein such as PD-1, which prevents or suppresses activation of cytotoxic T cells (CTLs) under normal circumstances. By inhibiting immunomodulatory proteins, for example, through the use of antibodies that link to these proteins, an increased T cell response to the tumor can be achieved. That is, these cancer immunoregulatory proteins suppress the immune response, and when this protein is blocked using, for example, an antibody against the immunoregulatory protein, the immune system is activated and induces immune stimulation, resulting in Provide treatment for conditions such as cancer and infectious diseases. Antibodies to either the PD-1 protein or its linking partner, PD-L1, induce T cell activation.

  Another area of current interest in combating disease utilizing the patient's immune system is T using agonist antibodies linked to costimulatory proteins such as ICOS (also known as inducible T cell costimulator, CD278). Includes cell costimulation, which adds a positive signal to overcome the negative signaling of immune checkpoint proteins on T cells. ICOS is a type I transmembrane protein containing an extracellular (Ig) V-like domain and functions as a receptor for B7h costimulatory molecules.

  Recent studies have shown that some tumor infiltrating lymphocytes (TIL) co-express PD-1 and ICOS (see Gros, J. Clinical Invest. 124 (5): 2246 (2014)). .

  Bispecific antibodies that can be linked to two different targets simultaneously offer the potential to improve the selectivity of TIL versus peripheral T cell targeting while simultaneously reducing therapeutic costs. A bivalent interaction between an antibody and two targets on the cell surface may in some cases lead to a higher linking affinity compared to a monovalent interaction with one target at a time. For this reason, normal bivalent antibodies tend to have a high linking affinity for their target on the cell surface. With bispecific antibodies, there is the potential to create higher selectivity for cells that simultaneously express two different targets, taking advantage of the higher affinity provided by simultaneous ligation to both targets.

  Accordingly, the present invention provides bispecific immunoregulatory antibodies that link to cells that express two antigens, as well as diseases such as cancer and infections by activating T cells and / or NK cells, and increased immune activity. Methods of treating other conditions that result in treatment are provided.

  Thus, the present invention provides bispecific antibodies that in some cases link to two different immunomodulatory molecules on a single cell, one a checkpoint receptor and the other a costimulatory receptor. And advantageously to solve the problem of toxicity and the cost of administering multiple antibodies by requiring administration of only a single therapeutic agent.

  Bispecific antibodies provide the opportunity to combine immunoimmunomodulation blockade with costimulation within one molecule. However, it is not clear which combination of immune regulation of immunity and costimulatory proteins, or which linkage stoichiometry (monovalent + monovalent, monovalent + divalent, etc.) is effective. Here, the present inventors can monovalently link to costimulatory proteins (such as ICOS) and monovalently link to checkpoint receptors (such as PD-1) to induce robust T cell activation. Identify bispecific antibodies.

  Surprisingly, although conventional knowledge states that monovalent antibodies do not cause agonism, this study shows the unexpected results of ICOS receptor agonism by monovalent bispecific antibodies of the invention. Merchant et al. , PNAS Jul 23 2013 E2987-E2996 “[w] hire initial screening of bibli- entialbodies produced agronists of MET, engineering the inventon tent. This positive result by only monovalent linking to ICOS requires at least bivalent linking to costimulatory proteins that provide the necessary level of receptor clustering on the cell surface to induce signal transduction. It is unexpected because it is considered to be. Fos et al. , J .; Immunol. ICOS ligation induces AKT phosphorylation, as shown by 2008: 1969-1977. The study here uses AKT phosphorylation as an indicator of ICOS agonism, and this effect is attributed to “one-arm ICOS” (see Example 5A (a)) and bispecific antibody monovalently linked to ICOS. Seen for both. As shown in Example 7 and FIG. 70, 1-arm XENP 20266, which connects only monovalently to ICOS, promotes more AKT phosphorylation than XENP16435 (eg, conventional mAb) which binds bivalently to ICOS. .

  Accordingly, the present invention relates to a novel construct for providing a heterodimeric bispecific antibody that allows linkage to a checkpoint receptor and human ICOS.

  In general, these bispecific antibodies are referred to as “anti-PD-1 × anti-ICOS”, or “PD-1 × ICOS” for each pair, generally for simplicity or simplicity (and thus interchangeable), etc. Named.

  The heterodimeric bispecific immunomodulatory antibodies of the invention are useful for treating various types of cancer. As will be appreciated by those skilled in the art, conventional monoclonal antibodies that link to tumor antigens, or newer classes of bispecific antibodies such as those linked to CD3 and tumor antigens (eg, as described in USSN 15 / 141,350). In contrast, immunomodulatory antibodies are used to enhance the immune response, but they are generally not tumor specific in their action. That is, the bispecific immunomodulatory antibodies of the invention inhibit suppression of the immune system and generally lead to T cell activation, which in turn leads to a greater immune response against cancerous cells and thus therapy.

  As discussed below, there are various ways in which T cell activation can be measured. The functional effect of bispecific immunomodulatory antibodies on NK cells and T cells should be assessed in vitro (and in some cases in vivo) by measuring changes in the following parameters: proliferation, cytokine release and cell surface markers Can do. With respect to NK cells, cell proliferation, cytotoxicity (ability to kill target cells as measured by increased CD107a, granzyme, and perforin expression, or by directly measuring target cell death), cytokines Increases in production (eg, IFN-γ and TNF), and expression of cell surface receptors (eg, CD25) are indicative of enhanced immune regulation, eg, death of cancer cells. For T cells, increased proliferation, cell surface markers of activation (eg, CD25, CD69, CD137, and PD1), cytotoxicity (ability to kill target cells), and cytokine production (eg, IL-2, IL -4, IL-6, IFN, TNF-α, IL-10, IL-17A) is an indication of enhanced immune regulation, for example, killing of cancer cells. Thus, treatment evaluation can be performed using an assay that evaluates one or more of the following: (i) increasing the immune response, (ii) αβ and / or γδ T cells. Increase activation, (iii) increase cytotoxic T cell activity, (iv) increase NK and / or NKT cell activity, (v) reduce αβ and / or γδ T cell suppression (Vi) increasing pro-inflammatory cytokine secretion; (vii) increasing IL-2 secretion; (viii) increasing interferon-γ production; (ix) increasing Th1 response; (X) reducing the Th2 response, (xi) reducing or eliminating at least one cell number and / or activity of regulatory T cells, (x i) IL-2 secretion to increase.

  Accordingly, in some embodiments, the present invention provides the use of a bispecific immunomodulatory antibody to perform one or more of the following in a subject in need thereof: (a) pro-inflammatory Up-regulating sex cytokines, (b) increasing T cell proliferation and / or increase, (c) increasing interferon-γ or TNF-α production by T cells, (d) IL-2 secretion (E) stimulating antibody response, (f) inhibiting cancer cell proliferation, (g) promoting antigen-specific T cell immunity, (h) CD4 + and / or CD8 + T cell activity (I) reduce T cell suppression, (j) promote NK cell activity, (k) promote apoptosis or lysis of cancer cells, and / or (L) Cytotoxic or cytostatic effect on cancer cells.

  Accordingly, the present invention provides bispecific immunomodulatory antibodies. There are several formats that can be used in the present invention, generally as shown in FIG. 2, many of which are heterodimers (such as, but not all, DVD-Ig).

  Heterodimeric antibody constructs are based on the self-assembly of two Fc domains of an antibody heavy chain, for example, two “monomers” that assemble into a “dimer”. Heterodimeric antibodies are made by modifying the amino acid sequence of each monomer as discussed more fully below. Thus, the present invention generally has different constant regions on each chain to promote heterodimer formation and / or to allow easy purification of heterodimers from homodimers. It relates to the generation of heterodimeric antibodies that can be co-linked to two antigens in several ways, depending on amino acid mutations.

  Accordingly, the present invention provides bispecific immunomodulatory antibodies. An ongoing problem in antibody technology is the need for “bispecific” antibodies that simultaneously link two (or more) different antigens, thereby generally bringing the different antigens in close proximity and new functionality and new therapies. Bring the law. Generally, these antibodies are made by including in the host cell a gene for each heavy and light chain (generally, in the present invention, two heavy chain monomers and a light chain as outlined herein). Gene for the chain). This generally results in the formation of the desired heterodimer (AB), as well as two homodimers (AA and BB). However, a major obstacle in the formation of bispecific antibodies is the purification of heterodimeric antibodies from homodimeric antibodies and / or bias towards heterodimer formation rather than homodimer formation. It is difficult to make it.

  To solve this problem, there are several mechanisms that can be used to generate the heterodimers of the present invention. Furthermore, as will be appreciated by those skilled in the art, these mechanisms can be combined to ensure high heterodimerization. Thus, amino acid mutations that result in the production of heterodimeric antibodies are referred to as “heterodimerization mutations”. As described below, heterodimerization mutations are steric mutations that allow purification of homodimers from heterodimers (eg, “knob and hole” or “skew” mutations described below and “charge pairs”). "Mutation") as well as "pI mutation".

  One mechanism is commonly referred to in the art as “knob and hole” (“KIH”) or often referred to herein as “skew” mutations, favoring heterodimer formation and homodimer formation. Refers to amino acid engineering that results in steric effects that can be disadvantageous, and can optionally be used and is often referred to as “knob and hole” and is described in Ridgway et al. , Protein Engineering 9 (7): 617 (1996), Atwell et al. , J .; Mol. Biol. 1997270: 26, US Pat. No. 8,216,805, US2012 / 0149876, all of which are hereby incorporated by reference in their entirety. These figures identify several “monomer A-monomer B” pairs that contain “knob and hole” amino acid substitutions. In addition, Merchant et al. , Nature Biotech. 16: 677 (1998), these “knob and hole” mutations can be combined with disulfide linkages for skew formation for heterodimerization. T366S / L368A / Y407V paired with T366W and T366S / L368A / Y407V / Y349C paired with T366W / S354C, this mutation with a bridged disulfide, specifically include other pI mutations as outlined below. In combination with the heterodimerization mutation of

  Additional mechanisms used to generate heterodimeric antibodies are described in Gunasekaran et al., Which is incorporated herein by reference in its entirety. , J .; Biol. Chem. 285 (25): 19637 (2010), often referred to as “electrostatic steering” or “charge pairs”. This is often referred to herein as a “charge pair”. In this embodiment, electrostatics is used to skew the formation towards heterodimerization. As will be appreciated by those skilled in the art, they can also affect pI and thus purification, and can therefore be considered as pI mutations in some cases. However, they are classified as “steric mutations” because they were generated to force heterodimerization and were not used as a purification tool. These include D221E / P228E / L368E paired with D221R / P228R / K409R (eg, these are “corresponding sets of monomers”), and C220E paired with C220R / E224R / P228R / K409R. / P228E / 368E, as well as others shown in the figures, but not limited to.

  In the present invention, in some embodiments, a pi mutation is used to modify the pi of one or both of the monomers, thus isoelectric of AA, AB and BB dimeric proteins. Allows point purification.

  In the present invention, there are several basic mechanisms that can facilitate the purification of heterodimeric proteins, one depending on the use of the pI mutation, such that each monomer has a different pI. Thus, isoelectric point purification of AA, AB and BB dimeric proteins becomes possible. Alternatively, some scaffold formats, such as the “Triple F” or “Bottle Opener” format, also allow separation based on size. It is also possible to “skew” the formation of a heterodimer rather than a homodimer, as further outlined below. Thus, combinations of steric heterodimerization mutations and pI or charge pair mutations are particularly used in the present invention. Furthermore, as outlined more fully below, scaffolds that utilize scFv, such as the Triple F format, can include a charged scFv linker (either positive or negative) that provides additional pI boost for purification purposes. As will be appreciated by those skilled in the art, the present invention also provides for the use of skew mutations with similarly charged scFv linkers (and combinations of Fc, FcRn, and KO mutations discussed herein) The triple F format is also useful when simply using a charged scFv linker and no further pI adjustment.

  In the present invention that utilizes pI as a separation mechanism to allow purification of heterodimeric proteins, amino acid mutations can be introduced into one or both monomeric polypeptides, ie, monomeric One (referred to herein as “monomer A” for simplicity) can be modified to differ from monomer B, or changes in both monomers A and B can be Modifications can be made to increase the pI of body A and decrease the pI of monomer B. As outlined more fully below, pi changes in either or both monomers can be achieved by removing or adding charged residues (eg, replacing neutral amino acids with positively or negatively charged amino acid residues). By changing the charged residue from a positive or negative to the opposite charge (eg from aspartic acid to lysine), or by changing the charged residue to a neutral residue (eg loss of charge, Lysine to serine). Some of these mutations are shown in the figure. Furthermore, pI mutations suitable for use in generating heterodimeric antibodies herein are those that are isotypes, eg, from different IgG isotypes such that the pi changes without introducing significant immunogenicity. See FIG. 29 of US Patent Application Publication No. 2014/288275, which is incorporated by reference herein in its entirety.

  Thus, this embodiment of the invention provides for causing sufficient pi changes in at least one monomer so that the heterodimer can be separated from the homodimer. As will be appreciated by those skilled in the art and as discussed further below, this is engineered to increase or decrease the “wild type” heavy chain constant region and its pI (wtA− + B or wtA−−B). Can be performed by using the mutated mutation region or by increasing one region and decreasing the other region (A + -B- or A-B +).

  Thus, in general, components of some embodiments of the present invention may include at least one isoelectric point (pI), if not both, of a monomer of a dimeric protein, an amino acid substitution (a “pI mutation”). Or “pI substitution”) by incorporation into one or both of the monomers to form “pI heterodimers” (when the protein is an antibody, these are called “pI antibodies”). Amino acid mutation in the constant region of the antibody of interest. As shown herein, separation of the heterodimer from the two homodimers can be achieved when the pi of the two monomers is slightly different from 0.1 pH units, 0.2 All that differ by 0.3, 0.4, and 0.5 or more are used in the present invention.

  As will be appreciated by those skilled in the art, the number of pi mutations to be included in each monomer or both monomers to obtain good separation depends in part on the desired scFv and the starting pi of the Fab. Will do. That is, to determine which monomer to manipulate or which “direction” (eg, more positive or more negative), the Fv sequences of the two target antigens are calculated and made from there . As is known in the art, different Fvs will have different starting pi utilized in the present invention. In general, as outlined herein, the pi is engineered to produce a total pi difference of at least about 0.1 log of each monomer, and as outlined herein, 0.2 to 0.5 is preferred. Further, as understood by those skilled in the art and as outlined herein, in some cases (depending on the format) the heterodimer may be separated from the homodimer based on size (eg, molecular weight). Can do. For example, as shown in some embodiments of FIG. 2, some formats result in different sized homodimers and heterodimers (eg, for a bottle opener, one homodimer is , "Dual scFv" format, one homodimer is a standard antibody, and a heterodimer has one Fab and one scFv).

  Design and purify multispecific proteins, including antibodies, when pi mutations are used to achieve a heterodimer that is purified over a homodimer by using the constant region of the heavy chain A more modular approach for providing is provided. Thus, in some embodiments, heterodimerization mutations (including skew and purified heterodimerization mutations) are not included in the variable region, so each individual antibody must be engineered. Further, in some embodiments, the potential for immunogenicity resulting from a pi mutation is achieved by transferring pi mutations from different IgG isotypes such that the pi changes without introducing significant immunogenicity. Remarkably reduced. Thus, a further problem to be solved is the elucidation of low pI constant domains with high human sequence content, eg minimization or avoidance of non-human residues at any particular position.

  Side effects that can occur with this pI engineering are also increased serum half-life and increased FcRn linkage. That is, reducing the pI of antibody constant domains (including those found in antibodies and Fc fusions) as described in USSN 13 / 194,904 (incorporated herein by reference in its entirety). May result in longer serum retention in vivo. These pi mutations to increase serum half-life also promote pi changes for purification.

  Furthermore, the pI mutation of the heterodimerization mutation has a significant ability to eliminate, minimize, and discriminate when homodimers are present, further adding to the bispecific antibody analysis and quality control process. Give profit. Similarly, the ability to reliably test the reproducibility of heterodimeric protein production is important.

  The first and second antigens of the present invention are referred to herein as antigen-1 and antigen-2, respectively, one is a costimulatory receptor and the other is a checkpoint receptor. One heterodimeric scaffold specifically used in the present invention is a “triple f” or “bottle opener” scaffold format. In this embodiment, one heavy chain of the antibody comprises a single chain fv (“scfv” as defined below) and the other heavy chain in a “normal” fab format comprising a variable heavy chain and a light chain. is there. This structure is often referred to as the “triple f” format (scfv-fab-fc) or “bottle opener” format because of the approximate similarity to the bottle opener (see FIG. 2). The two chains are brought together by the use of amino acid mutations in constant regions (eg, Fc domains and / or hinge regions) that facilitate heterodimeric antibody formation, as described more fully below.

  There are several distinct advantages to the current “Triple F” format. As is known in the art, antibody analogs that rely on two scFv constructs often have stability and aggregation problems, which in the present invention is the “normal” heavy and light chain pairing. It can be reduced by adding. In addition, there is a problem of mispairing heavy chain and light chain (eg heavy chain 1 paired with light chain 2) as opposed to a format that relies on two heavy chains and two light chains. do not do.

  In addition, as outlined herein, additional amino acid mutations can be introduced into the bispecific antibodies of the present invention to add additional functionality. For example, adding an amino acid change within the Fc region (to one monomer or both) to promote an increase in ADCC or CDC of the resulting molecule (eg, a change in linkage to the Fcγ receptor), as well as FcRn Can be increased and / or serum half-life can be increased. As described further herein and as will be appreciated by those skilled in the art, any and all of the mutations outlined herein may be arbitrarily and independently combined with other mutations. Can do.

  Similarly, another category of functional mutations are “Fcγ deletion mutations” or “Fc knockout (FcKO or KO) mutations. In these embodiments, for some therapeutic uses, avoid further mechanisms of action. In order to do so, it is desirable to reduce or eliminate normal linkage of the Fc domain to one or more or all Fcγ receptors (eg, FcγR1, FcγRIIa, FcγRIIb, FcγRIIIa, etc.), eg, removing FCγRIIIa linkage. It is generally desirable to eliminate or significantly reduce ADCC activity, with suitable ablation mutations shown in FIG.

IV. Definitions In order that the present invention may be more fully understood, some definitions are set forth below. Such definitions are intended to encompass grammatical equivalents.

  As used herein, “removal” means a reduction or removal of activity. Thus, for example, “removing an FcγR linkage” means that an Fc region amino acid mutation has less than 50% starting linkage compared to an Fc region that does not contain that particular mutation, A loss of activity of less than 90-95-98% is preferred and generally the activity is below the level of activity detectable in the Biacore assay. What is particularly useful in removing FcγR linkages is shown in FIG.

  As used herein, “ADCC” or “antibody-dependent cell-mediated cytotoxicity” refers to nonspecific cytotoxic cells that express FcγR recognize a linked antibody on a target cell, followed by the target cell A cell-mediated reaction that causes lysis of ADCC correlates with linkage to FcγRIIIa, and increased linkage to FcγRIIIa results in increased ADCC activity. As discussed herein, many embodiments of the invention completely eliminate ADCC activity.

  As used herein, “ADCP” or antibody-dependent cell-mediated phagocytosis refers to nonspecific cytotoxic cells expressing FcγR recognizing a linked antibody on a target cell, followed by phagocytosis of the target cell. Means a cell-mediated reaction that triggers.

As used herein, “antigen-binding domain” or “ABD” refers to six complementarity determining regions that, when present as part of a polypeptide sequence, specifically link to a target antigen as discussed herein. Means a set of (CDR). Thus, an “immunomodulating antigen-binding domain” is linked to a target immunomodulating antigen as outlined herein. As known in the art, these CDRs are as a first set of variable heavy chain CDRs (vh CDR or V _H CDR) and a second set of variable light chain CDRs (vl CDR or _VL CDR). Generally present, each contains three CDRs: vhCDR1, vhCDR2, vhCDR3 for the heavy chain and vlCDR1, vlCDR2 and vlCDR3 for the light chain. CDRs are present in variable heavy and variable light chain domains, respectively, and together form an Fv region. Thus, in some cases, the six CDRs of the antigen linking domain are contributed by the variable heavy and variable light chains. In the “Fab” format, a set of six CDRs includes two different polypeptide sequences: a variable heavy chain domain (including vh or V _H , vhCDR1, vhCDR2 and vhCDR3) and a variable light chain domain (vl or V _L , VlCDR1, vlCDR2 and vlCDR3), of which the C-terminus of the vh domain is linked to the N-terminus of the CH1 domain of the heavy chain and the C-terminus of the vl domain is linked to the N-terminus of the constant light chain domain (Thus forming a light chain). In the scFv format, the vh and vl domains are covalently linked to a single polypeptide sequence, generally by use of a linker as outlined herein (starting from the N-terminus) vh-linker- It can be either vl or vl-linker-vh, the former being generally preferred (including optional domain linkers on both sides, depending on the format used).

  As used herein, “modification” refers to amino acid substitutions, insertions, and / or deletions in a polypeptide sequence, or alterations to a moiety that is chemically linked to a protein. For example, the modification may be a modified carbohydrate or PEG structure linked to the protein. As used herein, “amino acid modification” refers to amino acid substitutions, insertions, and / or deletions in a polypeptide sequence. For clarity, unless otherwise specified, amino acid modifications are always for amino acids encoded by DNA, eg, 20 amino acids having codons in DNA and RNA.

  As used herein, “amino acid substitution” or “substitution” means substitution of an amino acid at a specific position in a parent polypeptide sequence with a different amino acid. In particular, in some embodiments, the substitution is for an amino acid that is not naturally occurring at a particular position (not naturally occurring in the organism or not present in any organism). For example, the substitution E272Y refers to a mutant polypeptide in which glutamic acid at position 272 is replaced with tyrosine, in this case an Fc mutation. For clarity, change the nucleic acid coding sequence but do not change the starting amino acid (eg, replace CGG (encodes arginine) with CGA (encodes arginine) to increase host organism expression levels) ) Is not an “amino acid substitution”. That is, if a protein has the same amino acid at the particular position where it begins, despite the creation of a new gene that encodes the same protein, it is not an amino acid substitution.

  “Amino acid insertion” or “insertion” as used herein means the addition of an amino acid sequence at a specific position in a parent polypeptide sequence. For example, -233E or 233E indicates the insertion of glutamic acid after position 233 and before position 234. Furthermore, -233ADE or A233ADE indicates insertion of AlaAspGlu after position 233 and before position 234.

  “Amino acid deletion” or “deletion” as used herein means the removal of an amino acid sequence at a specific position in a parent polypeptide sequence. For example, E233- or E233 #, E233 () or E233del indicates a deletion of glutamic acid at position 233. Furthermore, EDA233- or EDA233 # indicates a deletion of the sequence GluAspAla starting at position 233.

  As used herein, “mutant protein” or “protein variant” or “variant” means a protein that differs from that of the parent protein based on at least one amino acid modification. A protein variant may refer to the protein itself, a composition comprising the protein, or the amino sequence that encodes it. Preferably, the protein variant has at least one amino acid modification compared to the parent protein, for example from about 1 to about 70 amino acid modifications compared to the parent, preferably from about 1 to about 5 amino acid modifications. . As described below, in some embodiments, a parent polypeptide, eg, an Fc parent polypeptide, is a human wild-type sequence, such as an Fc region from IgG1, IgG2, IgG3, or IgG4, but has a mutation. Human sequences can also serve as “parent polypeptides” and can include, for example, IgG1 / 2 hybrids. The protein variant sequences herein preferably have at least about 80% identity, most preferably at least about 90% identity, more preferably at least about 95-98-99% identity with the parent protein sequence. Have sex. A mutant protein can refer to the mutant protein itself, a composition comprising a protein mutation, or a DNA sequence that encodes it.

  Thus, as used herein, “antibody variant” or “variant antibody” means an antibody that differs from a parent antibody based on at least one amino acid modification, and is used herein as “IgG “Variant” or “variant IgG” means an antibody that differs from a parent IgG (again, often derived from human IgG) based on at least one amino acid modification, and is used herein. By “immunoglobulin variant” or “variant immunoglobulin” is meant an immunoglobulin sequence that differs from that of the parent immunoglobulin sequence based on at least one amino acid modification. As used herein, “Fc variant” or “variant Fc” refers to a protein that contains amino acid modifications in the Fc domain. The Fc mutations of the present invention are defined according to the amino acid modifications that constitute them. Thus, for example, N434S or 434S is an Fc mutation with a substituted serine at position 434 relative to the parent Fc polypeptide, where numbering is according to the EU index. Similarly, M428L / N434S defines an Fc mutation with substitutions M428L and N434S relative to the parent Fc polypeptide. The identity of the WT amino acid may not be specified, in which case the mutation described above is called 428L / 434S. The order in which substitutions are provided is arbitrary, ie, for example, 428L / 434S is the same Fc mutation as M428L / N434S. For all positions discussed in the present invention relating to antibodies, the numbering of amino acid positions follows the EU index unless otherwise stated. EU index, or EU index similar to Kabat or EU numbering scheme, refers to EU antibody numbering (Edelman et al., 1969, Proc Natl Acad Sci USA 63: 78-85, which is hereby incorporated by reference in its entirety. Incorporated). The modification can be an addition, deletion, or substitution. Substitutions can include naturally occurring amino acids and optionally synthetic amino acids. Examples include US Pat. No. 6,586,207, WO 98/48032, WO 03/073238, US 2004-0214988A1, WO 05 / 35727A2, WO 05 / 74524A2, J. Pat. W. Chin et al. (2002), Journal of the American Chemical Society 124: 9026-9027; W. Chin, & P. G. Schultz, (2002), ChemBioChem 11: 1135-1137, J. MoI. W. Chin, et al. , (2002), PICAS United States of America 99: 11020-11024, and L.L. Wang, & P.W. G. Schultz, (2002), Chem. 1-10 are included, all incorporated by reference in their entirety.

  As used herein, “protein” as used herein means at least two covalently linked amino acids, including proteins, polypeptides, oligopeptides, and peptides. Peptidyl groups are naturally occurring amino acids and peptide linkages, or synthetic peptidomimetic structures, such as peptoids (see Simon et al., PNAS USA 89 (20): 9367 (1992), which is incorporated by reference in its entirety), etc. An “analog” may be included. The amino acid may be naturally occurring or synthetic (eg, not an amino acid encoded by DNA), as will be appreciated by those skilled in the art. For example, homophenylalanine, citrulline, ornithine, and noreoleucine are considered synthetic amino acids for the purposes of the present invention, and both D- and L- (R or S) configuration amino acids can be utilized. The mutations of the present invention are described, for example, in Cropp & Shultz, 2004, Trends Genet. 20 (12): 625-30, Anderson et al. , 2004, Proc Natl Acad Sci USA 101 (2): 7566-71, Zhang et al. , 2003, 303 (5656): 371-3, and Chin et al. , 2003, Science 301 (5635): 964-7, which may include modifications including the use of incorporated synthetic amino acids using techniques developed by Schultz and colleagues. In addition, polypeptides include synthetic derivatization of one or more side chains or termini, glycosylation, PEGylation, cyclic sequences, cyclization, linkers to other molecules, fusions to proteins or protein domains, and peptide tags Or the addition of a label may be included.

  “Residue” as used herein means a position in a protein and its associated amino acid identity. For example, asparagine 297 (also referred to as Asn297 or N297) is the residue at position 297 in the human antibody IgG1.

  As used herein, “Fab” or “Fab region” means a polypeptide comprising VH, CH1, VL, and CL immunoglobulin domains. Fab may refer to this region in isolation, or this region associated with a full-length antibody, antibody fragment, or Fab fusion protein.

  “Fv” or “Fv fragment” or “Fv region” as used herein means a polypeptide comprising the VL and VH domains of a single ABD. As will be appreciated by those skilled in the art, these can generally be composed of two chains or combined (generally by linkers as discussed herein) to form scFvs. In some cases, for example, “center-Fv” and “DVD-Ig” formats add “extra” vh and vl domains that function as scFv, but the vh and vl domains are linked using an scFv linker. Not.

  As used herein, a “single chain Fv” or “scFv” is a variable that is covalently linked to a variable light chain domain, generally using an scFv linker as discussed herein, to form an scFv or scFv domain. Refers to the heavy chain domain. The scFv domain can be in any orientation from N-terminal to C-terminal (vh-linker-vl or vl-linker-vh).

  As used herein, “IgG subclass modification” or “isotype modification” means an amino acid modification that converts one amino acid of one IgG isotype to the corresponding amino acid of a different, matched IgG isotype. . For example, at EU position 296, IgG1 contains tyrosine and IgG2 contains phenylalanine, so the F296Y substitution in IgG2 is considered an IgG subclass modification.

  As used herein, “non-naturally occurring modifications” means amino acid modifications that are not isotypes. For example, since none of IgG contains a serine at position 434, substitution 434S in IgG1, IgG2, IgG3, or IgG4 (or hybrids thereof) is considered a non-naturally occurring modification.

  As used herein, “amino acid” and “amino acid identity” means one of the 20 naturally occurring amino acids encoded by DNA and RNA.

  As used herein, “effector function” means a biochemical event that results from the interaction of an antibody Fc region with an Fc receptor or ligand. Effector functions include, but are not limited to, ADCC, ADCP, and CDC.

  As used herein, “IgG Fc ligand” means a molecule, preferably a polypeptide, from any organism that is linked to the Fc region of an IgG antibody to form an Fc / Fc ligand complex. Fc ligands include, but are not limited to, FcγRI, FcγRII, FcγRIII, FcRn, C1q, C3, mannan-linked lectin, mannose receptor, staphylococcal protein A, streptococcal protein G, and viral FcγR. Fc ligands also include the Fc receptor homologue (FcRH), a family of Fc receptors that are homologous to FcγR (Davis et al., 2002, Immunological Reviews 190: 123-136, incorporated by reference in its entirety. ) Fc ligands can include undiscovered molecules that link to Fc. Particular IgG Fc ligands are FcRn and Fc gamma receptors. As used herein, “Fc ligand” means a molecule, preferably a polypeptide, from any organism that is linked to the Fc region of an antibody to form an Fc / Fc ligand complex.

  As used herein, “Fc gamma receptor”, “FcγR”, or “Fc gamma R” refers to any member of the protein family linked to the IgG antibody Fc region and encoded by the FcγR gene. Means. In humans, this family includes FcγRI (CD64) including isoforms FcγRIa, FcγRIb, and FcγRIc; FcγRIIa (including allotypes H131 and R131), FcγRIIb (FcγRIIb-1 and FcγRIIb-2) As well as FcγRIIc containing FcγRIIc (CD32); and FcγRIIIa (including allotypes V158 and F158) and FcγRIIIb (including allotypes FcγRIIb-NA1 and FcγRIIb-NA2) CD16) (Jefferis et al., 2002, Immunol Lett 82: 57-65, which is incorporated by reference in its entirety). As well as it includes any undiscovered human FcγR or FcγR isoforms or allotypes, without limitation. The FcγR can be from any organism, including but not limited to humans, mice, rats, rabbits, and monkeys. Mouse FcγR includes FcγRI (CD64), FcγRII (CD32), FcγRIII-1 (CD16), and FcγRIII-2 (CD16-2), and any undiscovered mouse FcγR or FcγR isoform or allotype However, it is not limited to these.

  As used herein, “FcRn” or “neonatal Fc receptor” refers to a protein linked to the Fc region of an IgG antibody and encoded at least in part by the FcRn gene. The FcRn can be derived from any organism, including but not limited to humans, mice, rats, rabbits, and monkeys. As is known in the art, a functional FcRn protein comprises two polypeptides, often called heavy and light chains. The light chain is β-2-microglobulin and the heavy chain is encoded by the FcRn gene. Unless stated otherwise herein, FcRn or FcRn protein refers to a complex of FcRn heavy chain and β-2-microglobulin. Various FcRn mutations used to increase linkage to the FcRn receptor and in some cases to increase serum half-life are shown in the figure legend of FIG.

  As used herein, “parent polypeptide” refers to a starting polypeptide that is subsequently modified to produce variants. The parent polypeptide may be a naturally occurring polypeptide, or a variant or engineered version of a naturally occurring polypeptide. The parent polypeptide may refer to the polypeptide itself, a composition comprising the parent polypeptide, or the amino acid sequence that encodes it. Thus, “parent immunoglobulin” as used herein refers to an unmodified immunoglobulin polypeptide that is modified to produce variants, and “parent antibody” as used herein is a variant. An unmodified antibody that is modified to produce a type antibody. It should be noted that “parent antibody” includes known commercially available recombinantly produced antibodies, as outlined below.

  As used herein, “Fc” or “Fc region” or “Fc domain” means a constant region of an antibody, excluding a first constant region immunoglobulin domain, and optionally a portion of a hinge. Means peptide. Thus, Fc refers to the last two constant region immunoglobulin domains of IgA, IgD, and IgG, the last three constant region immunoglobulin domains of IgE and IgM, and the mobile hinge N-terminus to these domains. For IgA and IgM, Fc may include the J chain. In the case of IgG, the Fc domain includes immunoglobulin domains Cγ2 and Cγ3 (Cγ2 and Cγ3) and a lower hinge region between Cγ1 (Cγ1) and Cγ2 (Cγ2). Although the boundaries of the Fc region may vary, a human IgG heavy chain Fc region is usually defined as containing residues C226 or P230 at its carboxy terminus, where the numbering follows a EU index similar to Kabat. In some embodiments, amino acid modifications are made to the Fc region, eg, to alter linkage to one or more FcγR or FcRn receptors, as described in more detail below. .

  As used herein, “heavy chain constant region” refers to the CH1-hinge-CH2-CH3 portion of an antibody.

  As used herein, an “Fc fusion protein” or “immunoadhesin” generally refers to a different protein, eg, as described herein (optionally via a linker moiety as described herein). It means a protein containing an Fc region linked to a linking moiety to a target protein. In some cases, one monomer of a heterodimeric antibody comprises an antibody heavy chain (contains an scFv or further comprises a light chain) and the other monomer comprises an Fc domain comprising a variant Fc domain and a ligand. Is the body. In some embodiments, these “half antibody-half fusion proteins” are referred to as “fusions”.

  As used herein, “position” means a location in the sequence of a protein. The positions may be numbered sequentially or according to an established format, eg, an EU index for antibody numbering.

  “Target antigen” as used herein means a molecule to which the variable regions of a given antibody are specifically linked. The target antigen may be a protein, carbohydrate, lipid, or other compound. Suitable target antigens are described below.

  “Strandedness” in conjunction with the heterodimeric antibody monomers of the invention herein is “matched” to the heterodimer, as is the “matched” double-stranded DNA. It means that heterodimerization mutations are incorporated into each monomer so as to retain the ability to form bodies. For example, if several pI mutations are engineered into monomer A (eg, making pI higher), a stereo charge that is a “charge pair” that can also be utilized does not interfere with the pI mutation, eg , Charge mutations with increased pI are placed in the same “strand” or “monomer” and retain both functions. Similarly, as outlined in more detail below, for “skew” mutations that result in a paired set, one of skill in the art will determine in which chain or monomer that incorporates one set of pairs will fall. Consider the pI and use the skew pI as well to maximize the pI separation.

  “Target cell” as used herein means a cell that expresses a target antigen.

  A “variable region” as used herein is one or more Ig domains substantially encoded by either Vκ (V. kappa) or Vλ (V. lambda), and / or each , And kappa, lambda, and immunoglobulin regions, including the VH genes that make up the heavy chain immunoglobulin loci.

  As used herein, “wild type or WT” means an amino acid or nucleotide sequence found in nature, including allelic variations. A WT protein has an amino acid or nucleotide sequence that is not intentionally modified.

  The antibodies of the invention are generally isolated or recombinant. As used to describe the various polypeptides disclosed herein, “isolated” is identified and separated and / or recovered from the cell or cell culture in which it was expressed. Means polypeptide. Ordinarily, an isolated polypeptide is prepared by at least one purification step. An “isolated antibody” refers to an antibody that is substantially free of other antibodies with different antigen specificities. “Recombinant” means that the antibody is produced in a foreign host cell using recombinant nucleic acid technology.

  “Percent amino acid sequence identity (%)” for a protein sequence is the amino acid residue in a particular (parent) sequence after aligning the sequence to achieve the maximum percent sequence identity and introducing gaps if necessary. Defined as the percentage of amino acid residues in the candidate sequence that are identical to the group, and does not consider any conservative substitutions as part of the sequence identity. Alignments for the purpose of determining percent amino acid sequence identity are within the skill of the art using publicly available computer software such as BLAST, BLAST-2, ALIGN, or Megalign (DNASTAR) software. Can be achieved in various ways. One skilled in the art can determine appropriate parameters for measuring alignment, including any algorithms needed to achieve maximal alignment over the full length of the sequences being compared. One particular program is the ALIGN-2 program outlined in US Patent Application Publication No. 2016/0244525, paragraphs [0279]-[0280], which is incorporated herein by reference.

  The degree of identity between the amino acid sequence of the present invention (the “sequence of the present invention”) and the parent amino acid sequence is the shorter of the length of the “sequence of the present invention” or the length of the parent sequence. Calculated as the number of exact matches in the alignment of the two sequences divided by. Results are expressed as percent identity.

  In some embodiments, the two or more amino acid sequences are at least 50%, 60%, 70%, 80%, or 90% identical. In some embodiments, the two or more amino acid sequences are at least 95%, 97%, 98%, 99%, or even 100% identical.

  “Specific linkage” or “specifically linked to” or “specific for” to a specific antigen or epitope means a linkage that is measurablely different from a non-specific interaction. To do. Specific ligation can be measured, for example, by determining the ligation of a molecule relative to that of a control molecule, which is generally a molecule of similar structure that does not have ligation activity. For example, specific linkage can be determined by competition between a target and a similar control molecule.

Specific linkage to a particular antigen or epitope is, for example, at least about 10 ⁻⁴ M, at least about 10 ⁻⁵ M, at least about 10 ⁻⁶ M, at least about 10 ⁻⁷ M, at least about 10 for the antigen or epitope. ⁻⁸ M, at least about 10 ⁻⁹ M, alternatively at least about 10 ⁻¹⁰ M, at least about 10 ⁻¹¹ M, at least about 10 ⁻¹² , or more can be indicated by an antibody having a KD, Refers to the dissociation rate of a particular antibody-antigen interaction. Typically, an antibody that specifically binds to an antigen is 20-fold, 50-fold, 100-fold, 500-fold, 1000-fold, 5,000-fold, 10,000, or 10,000 relative to a control molecule relative to an antigen or epitope. You will have a KD that is twice or greater.

  In addition, specific linking to a specific antigen or epitope is such that, for example, the KA or Ka for the antigen or epitope is at least 20 times, 50 times, 100 times, 500 times, 1000 times that epitope relative to the control. , 5,000-fold, 10,000-fold, or more, may be indicated by an antibody, where KA or Ka refers to the association rate of a particular antibody-antigen interaction. Ligation affinity is generally measured using a Biacore assay.

V. Antibodies The present invention relates to the generation of bispecific immunomodulatory antibodies that link to two different immunomodulating antigens, as discussed herein. As discussed below, the term “antibody” is used generically. The antibodies used in the present invention can take several formats as described herein.

  Conventional antibody structural units typically comprise a tetramer. Each tetramer typically consists of two identical polypeptide chain pairs, each pair consisting of one “light” chain (typically a molecular weight of about 25 kDa) and one With a “heavy” chain (typically a molecular weight of about 50-70 kDa). Human light chains are classified as kappa light chains and lambda light chains. The present invention is generally directed to bispecific antibodies based on the IgG class, which has several subclasses, including but not limited to IgG1, IgG2, IgG3, and IgG4. In general, IgG1, IgG2, and IgG4 are used more frequently than IgG3. It should be noted that IgG1 has different allotypes with polymorphisms at 356 (D or E) and 358 (L or M). The sequences illustrated herein use the 356D / 358M allotype, although other allotypes are included herein. That is, any sequence comprising an IgG1 Fc domain included herein can have 356E / 358L instead of the 356D / 358M allotype.

  Further, many of the sequences herein are those in which at least one cysteine at position 220 is replaced with serine, which is generally on the “scFv monomer” side for most of the sequences described herein. However, it can be on the “Fab monomer” side or both to reduce disulfide formation. Specifically included within the sequences herein are those in which one or both of these cysteines are substituted (C220S).

  Thus, as used herein, “isotype” refers to any subclass of immunoglobulin defined by the chemical and antigenic characteristics of their constant regions. It should be understood that therapeutic antibodies may also include isotype and / or subclass hybrids. For example, the present invention encompasses pI engineering of IgG1 / G2 hybrids, as shown in US Patent Application Publication No. 2009/0163699, which is incorporated herein by reference.

  The amino-terminal portion of each chain is a variable region of about 100-110 or more amino acids primarily involved in antigen recognition, commonly referred to in the art and herein as “Fv domains” or “Fv regions”. including. In the variable region, three loops are aggregated for each of the heavy and light chain V domains to form the antigen linking site. Each of the loops is referred to as a complementarity determining region (hereinafter referred to as “CDR”), where the variation in amino acid sequence is most prominent. “Variable” refers to the fact that certain segments of the variable region vary greatly in sequence between antibodies. The variability within the variable region is not evenly distributed. Instead, the V region is a 15-30 amino acid framework region (FR) separated by shorter regions of hypervariability, each termed a “hypervariable region” that is 9-15 amino acids long or longer. ) Is a relatively invariant section.

  Each VH and VL consists of three hypervariable regions ("complementarity determining regions", "CDRs") arranged in the order FR1-CDR1-FR2-CDR2-FR3-CDR3-FR4 from the amino terminus to the carboxy terminus, and Consists of four FRs.

  The hypervariable region is generally approximately the amino acid residues 24-34 (LCDR1, “L” represents the light chain), 50-56 (LCDR2), and 89-97 (LCDR3) in the light chain variable region and heavy Approximately 31-35B (HCDR1, “H” represents heavy chain), 50-65 (HCDR2), and 95-102 (HCDR3) amino acid residues (Kabat et al., SEQUENCES OF) in the chain variable region. PROTEINS OF IMMUNOLOGICAL INTEREST, 5th Ed.Public Health Service, National Institutes of Health, Bethesda, Md. (1991)) and / or residues that form hypervariable loops (eg, light chain variable regions) Residues 26-32 (LCDR1), 50-52 (LCDR2), and 91-96 (LCDR3) and 26-32 (HCDR1), 53-55 (HCDR2), and 96-101 in the heavy chain variable region HCDR3) (Cothia and Lesk (1987) J. Mol. Biol. 196: 901-917.) Specific CDRs of the invention are described below.

  As will be appreciated by those skilled in the art, the exact numbering and placement of CDRs can vary between different numbering systems. However, it should be understood that disclosure of variable heavy chains and / or variable light chains includes disclosure of related (inherent) CDRs. Thus, the disclosure of each variable heavy chain region is the disclosure of vhCDR (eg, vhCDR1, vhCDR2, and vhCDR3), and the disclosure of each variable light chain region is the disclosure of vlCDR (eg, vlCDR1, vlCDR2, and vlCDR3). is there.

A useful comparison of CDR numbering is as follows and is described by Lafranc et al. Dev. Comp. Immunol. 27 (1): 55-77 (2003). :

  Throughout this specification, the Kabat numbering system is generally used when referring to residues within the variable domain (approximately residues 1 to 107 of the light chain variable region and residues 1 to 113 of the heavy chain variable region). The EU numbering system is for the Fc region (see, eg, Kabat et al., Supra (1991)).

  The present invention provides a number of different CDR sets. In this case, the “complete CDR set” comprises three variable light chains and three variable heavy chain CDRs, for example vlCDR1, vlCDR2, vlCDR3, vhCDR1, vhCDR2, and vhCDR3. These can each be part of a larger variable light chain or variable heavy chain domain. Further, as outlined more fully herein, the variable heavy and light chain domains are distinct when heavy and light chains are used (eg, when Fab is used). It can be on a polypeptide chain or in the case of scFv sequences on a single polypeptide chain.

  CDRs contribute to the formation of antigen linkages, or more specifically to the formation of epitope-binding sites of antibodies. “Epitope” refers to a determinant that interacts with a specific antigen-binding site in the variable region of an antibody molecule known as a paratope. Epitopes are a classification of molecules such as amino acids or sugar side chains and usually have specific structural characteristics and specific charge characteristics. A single antigen may have more than one epitope.

  Epitopes are amino acid residues that are directly involved in linking (also referred to as the immunodominant component of the epitope) and other amino acid residues that are not directly involved in linking, eg, amino acids that are effectively blocked by specific antigen-linking peptides. Residues, in other words, amino acid residues that are within the footprint of the specific antigen-linked peptide.

  Epitopes may be either conformational or linear. Conformational epitopes are created by the spatial alignment of amino acids from different segments of a linear polypeptide chain. A linear epitope is one that is created by adjacent amino acid residues in a polypeptide chain. Conformational and non-conformational epitopes can be distinguished in that the connection to the former is lost in the presence of a denaturing solvent, but the connection to the latter is not lost.

  An epitope typically includes at least 3, and more usually, at least 5, or 8-10 amino acids in a unique spatial conformation. Antibodies that recognize the same epitope can be validated in a simple immunoassay that demonstrates the ability to block the linkage of one antibody to the target antigen of another antibody, eg, “binning”. As outlined below, the present invention includes not only the listed antigen-binding domains and the antibodies herein, but also those that compete for ligation with the epitopes linked by the listed antigen-binding domains.

  The carboxy-terminal portion of each chain defines a constant region primarily responsible for effector function. Kabat et al. Collected a number of primary sequences of heavy and light chain variable regions. Based on the degree of sequence conservation, Kabat et al. Classified individual primary sequences into CDRs and frameworks and created a list of them (SEQUENCES OF IMMUNOLOGICAL INTEREST, 5th edition, NIH publication, No. 1, incorporated by reference in its entirety). 91-3242, EA Kabat et al.).

  The IgG subclass of immunoglobulins has several immunoglobulin domains in the heavy chain. By “immunoglobulin (Ig) domain” herein is meant a region of an immunoglobulin that has a distinctly different tertiary structure. A heavy chain domain comprising a constant heavy chain (CH) domain and a hinge domain is of interest in the present invention. In connection with IgG antibodies, each IgG isotype has three CH regions. Thus, in connection with IgG, the “CH” domain is as follows: “CH1” refers to the 118th to 220th positions according to the EU index similar to Kabat. “CH2” refers to positions 237 to 340 according to the EU index similar to Kabat, and “CH3” refers to positions 341 to 447 according to the EU index similar to Kabat. As shown herein and described below, the pI mutation may be present in one or more of the CH regions and, as discussed below, the hinge region.

  Another type of heavy chain Ig domain is the hinge region. As used herein, “hinge” or “hinge region” or “antibody hinge region” or “immunoglobulin hinge region” means a mobile polypeptide comprising amino acids between the first and second constant domains of an antibody. To do. Structurally, the IgG CH1 domain ends at EU position 220 and the IgG CH2 domain begins at residue EU position 237. Thus, for IgG, the antibody hinge is defined herein to include positions 221 (D221 in IgG1) to 236 (G236 in IgG1), and the numbering follows the EU index as in Kabat. In some embodiments, in conjunction with the Fc region, a lower hinge is included, and “lower hinge” generally refers to position 226 or 230. As described herein, pI mutations can be made in the hinge region as well.

  The light chain generally has two domains: a variable light chain domain (which includes the light chain CDRs and together with the variable heavy chain domain forms an Fv region), and a constant light chain region (often referred to as CL or Cκ). including.

  Another region of interest for additional substitutions outlined below is the Fc region.

  Thus, the present invention provides different antibody domains. As described herein and known in the art, the heterodimeric antibodies of the invention comprise different domains within the heavy and light chains, which may also overlap. These domains are: Fc domain, CH1 domain, CH2 domain, CH3 domain, hinge domain, heavy chain constant domain (CH1-hinge-Fc domain or CH1-hinge-CH2-CH3), variable heavy chain domain, variable light chain domain , Light chain constant domains, Fab domains, and scFv domains.

  Thus, an “Fc domain” includes a —CH 2 —CH 3 domain and optionally a hinge domain. In embodiments herein, when the scFv is linked to an Fc domain, it is the C-terminus of the scFv construct that is linked to all or part of the hinge of the Fc domain, eg, it is generally the beginning of the hinge. Is linked to the sequence EPKS. The heavy chain comprises a variable heavy chain domain and a constant domain, which comprises an Fc domain comprising CH1-optional hinge-CH2-CH3. The light chain includes a variable light chain and a light chain constant domain. The scFv includes a variable heavy chain, an scFv linker, and a variable light chain domain. In most of the constructs and sequences outlined herein, the C-terminus of the variable light chain is linked to the N-terminus of the scFv linker, which is linked to the N-terminus of the variable heavy chain (N-vh-linker). -Vl-C), but this can be switched (N-vl-linker-vh-C). Some embodiments of the invention comprise at least one scFv domain that does not exist in nature but generally comprises a variable heavy chain domain and a variable light chain domain joined together by an scFv linker. As outlined herein, scFv domains are generally oriented from the N-terminus to the C-terminus as VH-scFv linker-VL, which is a scFv domain (or Fab-derived VH and VL sequences). Can be reversed to the VL-scFv linker-VH using any linker at one or both ends, depending on the format (generally figured out). 2).

  As demonstrated herein, there are a number of suitable scFv linkers that can be used, including conventional peptide linkages produced by recombinant techniques. The linker peptide may mainly comprise the following amino acid residues: Gly, Ser, Ala, or Thr. The linker peptides should be long enough to link the two molecules in such a way that they are in the correct conformation with respect to each other so that they retain the desired activity. In one embodiment, the linker is about 1-50 amino acids long, preferably about 1-30 amino acids long. In one embodiment, a linker that is 1-20 amino acids in length may be used, and in some embodiments, about 5 to about 10 amino acids are useful. Useful linkers include, for example, (GS) n, (GSGGS) n, (GGGGGS) n, and (GGGS) n, where n is an integer of at least 1 (and generally 3-4). Polymers, glycine-alanine polymers, alanine-serine polymers, as well as other mobile linkers are included. Alternatively, a variety of non-proteinaceous polymers may be useful as linkers, including but not limited to polyethylene glycol (PEG), polypropylene glycol, polyoxyalkylene, or copolymers of polyethylene glycol and polypropylene glycol. It can be useful as a linker.

  Other linker sequences include any sequence of the CL / CH1 domain of any length, but may not include all residues of the CL / CH1 domain, eg, the first 5 to 5 of the CL / CH1 domain 12 amino acid residues. The linker may be derived from an immunoglobulin light chain, such as Cκ or Cλ. The linker can be derived from any isotype of immunoglobulin heavy chain including, for example, Cγ1, Cγ2, Cγ3, Cγ4, Cα1, Cα2, Cδ, Cε, and Cμ. The linker sequence may also be derived from other proteins such as Ig-like proteins (eg, TCR, FcR, KIR), hinge region derived sequences, and other native sequences from other proteins.

  In some embodiments, the linker is a “domain linker” used to link together any two domains outlined herein. Although any suitable linker can be used, many embodiments, for example, (GS) n, (GSSGGS) n, (GGGGGS), where n is at least 1 (and generally 3-4-5). glycine-serine polymers containing n, and (GGGS) n, and any that allows recombination of two domains with sufficient length and flexibility so that each domain can retain its biological function Peptide sequences are utilized. In some cases, as outlined below, charged domain linkers, such as those used in some embodiments of scFv linkers, can be used with attention to “twist”.

  In some embodiments, the scFv linker is a charged scFv linker, some of which are shown in FIG. Thus, the present invention further provides a charged scFv linker to facilitate the separation of pi between the first monomer and the second monomer. That is, by incorporating a positive or negative charged scFv linker (or both in the case of a scaffold using scFv on a different monomer), this allows for a single monomer containing the charged linker without further modification of the Fc domain. Makes it possible to change the pI of the body. These charged linkers can be substituted into any scFv including standard linkers. Also, as will be appreciated by those skilled in the art, charged scFv linkers are used on the correct “chain” or monomer according to the desired change in pI. For example, as discussed herein, to generate a triple F format heterodimeric antibody, the original pI of the Fv region for each of the desired antigen-binding domains is calculated and an scFv is generated. One is selected for, and either positive or negative linker is selected depending on the pI.

  Charged domain linkers can also be used to increase the pi separation of the monomers of the invention, and thus what is included in FIG. 8 is any implementation herein where a linker is utilized. Can be used in form

  In one embodiment, an antibody is an antibody fragment so long as it contains at least one constant domain that can be engineered to produce heterodimers such as pI engineering. Other antibody fragments that can be used include those comprising one or more of the CH1, CH2, CH3, hinge and CL domains of the present invention that have been pI engineered. In particular, the format shown in FIG. 1 is an antibody, commonly referred to as a “heterodimeric antibody”, in which the protein is at least two related Fc sequences self-assembled into a heterodimeric Fc domain, and as a Fab or scFv. That means having at least two Fv regions.

A. Chimeric and Humanized Antibodies In some embodiments, the antibodies herein can be derived from a mixture from different species, such as chimeric and / or humanized antibodies. In general, both “chimeric antibodies” and “humanized antibodies” refer to antibodies that combine regions from two or more species. For example, a “chimeric antibody” conventionally comprises a variable region derived from a mouse (or rat in some cases) and a constant region derived from a human. “Humanized antibody” generally refers to non-human antibodies in which variable domain framework regions are swapped with sequences found in human antibodies. In general, in humanized antibodies, the entire antibody, except the CDRs, is encoded by a polynucleotide of human origin or is identical to such an antibody except within its CDRs. CDRs partially or fully encoded by nucleic acids originating from non-human organisms are grafted into the beta sheet framework of the human antibody variable region to produce an antibody whose specificity is determined by the grafted CDR. . The production of such antibodies is described, for example, in WO 92/11018, Jones, 1986, Nature 321: 522-525, Verhoeyen et al. , 1988, Science 239: 1534-1536, all incorporated by reference in their entirety. “Backmutation” of selected acceptor framework residues to the corresponding donor residues is often required to restore the lost affinity in the initial transplanted construct (US 5530101, US 5585089). US Pat. No. 5,693,761, US Pat. No. 5,693,762, US Pat. No. 6,180,370, US Pat. No. 5,859,205, US Pat. No. 5,821,337, US Pat. No. 6,054,297, US Pat. A humanized antibody will also optimally comprise at least a portion, usually all, of an immunoglobulin constant region, typically that of a human immunoglobulin, and thus typically will comprise a human Fc region. . Humanized antibodies can also be generated using mice with a genetically engineered immune system. Roque et al. , 2004, Biotechnol. 20: 639-654, incorporated by reference in its entirety. Various techniques and methods for humanizing and reforming non-human antibodies are well known in the art (Tsurushita & Vasquez, 2004, Humanization of Monoclonal Antibodies, Molecular Biology of B Cells, 533-erse, 543-erse. (USA), and references cited therein, all incorporated by reference in their entirety). Humanization methods include Jones et al. , 1986, Nature 321: 522-525; Riechmann et al. , 1988; Nature 332: 323-329; Verhoeyen et al. , 1988, Science, 239: 1534-1536; Queen et al. 1989, Proc Natl Acad Sci, USA 86: 10029-33; He et al. 1998, J. MoI. Immunol. 160: 1029-1035; Carter et al. , 1992, Proc Natl Acad Sci USA 89: 4285-9, Presta et al. , 1997, Cancer Res. 57 (20): 4593-9; Gorman et al. 1991, Proc. Natl. Acad. Sci. USA 88: 4181-4185; O'Connor et al. 1998, Protein Eng 11: 321-8, including but not limited to, the entirety of which is incorporated by reference. Other methods of reducing the immunogenicity of humanized or non-human antibody variable regions include, for example, Roguska et al. , 1994, Proc. Natl. Acad. Sci. A surface reconstruction method as described in USA 91: 969-973 may be included.

  In certain embodiments, an antibody of the invention comprises a heavy chain variable region derived from a specific germline heavy chain immunoglobulin gene and / or a light chain variable region derived from a specific germline light chain immunoglobulin gene. . For example, such an antibody can comprise or consist of a human antibody comprising a heavy or light chain variable region that is “product of” or “derived from” a particular germline sequence. A human antibody that is “product of” or derived from a human germline immunoglobulin sequence comprises comparing the amino acid sequence of a human antibody with the amino acid sequence of a human germline immunoglobulin; By selecting the human germline immunoglobulin sequence that is the closest (ie, the highest% identity) can be identified as such. A human antibody that is “product” or “derived from” a particular human germline immunoglobulin sequence can be produced by, for example, deliberate introduction of naturally occurring somatic mutations or site-specific mutations. It may contain amino acid differences compared to the cell line sequence. However, humanized antibodies are typically at least 90% identical in amino acid sequence to the amino acid sequence encoded by a human germline immunoglobulin gene, and other species of germline immunoglobulin amino acid sequences (eg, Amino acid residues that identify the antibody as being derived from a human sequence when compared to the mouse germline sequence). In certain cases, a humanized antibody has an amino acid sequence encoded by a germline immunoglobulin gene and at least 95, 96, 97, 98 or 99% in amino acid sequence, or at least 96%, 97%, 98 %, Or even 99% may be the same. Typically, a humanized antibody derived from a particular human germline sequence will show no more than 10-20 amino acid differences from the amino acid sequence encoded by the human germline immunoglobulin gene (see herein). Before the introduction of any skew, pI and excision mutations, ie before the introduction of the mutation of the invention, the number of mutations is generally small). In certain cases, a humanized antibody may exhibit no more than 5 amino acid differences, or even no more than 4, 3, 2, or 1 amino acid difference from the amino acid sequence encoded by the germline immunoglobulin gene. (Also, the number of mutations is generally low before the introduction of any skew, pI, and excision mutations herein, ie, before the introduction of the mutations of the invention).

  In one embodiment, the parent antibody is affinity matured as is known in the art. Structure-based methods can be used for humanization and affinity maturation, for example, as described in USSN 11 / 004,590. Wu et al. 1999, J. MoI. Mol. Biol. 294: 151-162; Baca et al. , 1997, J. Am. Biol. Chem. 272 (16): 10678-10684; Rosok et al. 1996, J. MoI. Biol. Chem. 271 (37): 22611-22618; Rader et al. 1998, Proc. Natl. Acad. Sci. USA 95: 8910-8915; Krauss et al. , 2003, Protein Engineering 16 (10): 753-759, methods based on selection can be used for humanization and / or affinity maturation of antibody variable regions, including, but not limited to, All are hereby incorporated by reference. Other humanization methods may include transplanting only the portion of the CDR, which is fully incorporated by reference, USSN 09 / 810,510; Tan et al. , 2002, J. et al. Immunol. 169: 1119-1125; De Pascalis et al. , 2002, J. et al. Immunol. 169: 3076-3084, but is not limited to these.

VI. Accordingly, in some embodiments, the present invention provides for the use of two different heavy chain variant Fc sequences that self-assemble to form a heterodimeric Fc domain and a heterodimeric antibody. Dependent heterodimeric immunomodulatory antibodies are provided.

  The present invention relates to novel constructs for providing heterodimeric antibodies that allow linkage to more than one immunomodulatory antigen or ligand, eg, to allow bispecific linkage. Heterodimeric antibody constructs are based on the self-assembling nature of two “monomers” assembled into two Fc domains of an antibody heavy chain, eg, “dimers”. Heterodimeric antibodies are made by changing the amino acid sequence of each monomer as discussed more fully below. Thus, the present invention generally provides amino acid mutations in different constant regions on each chain to promote heterodimer formation and / or to facilitate the purification of heterodimers over homodimers. Depending on the generation of heterodimeric immunomodulatory antibodies that can co-link antigens in several ways.

  Accordingly, the present invention provides bispecific antibodies. An ongoing problem in antibody technology is the need for “bispecific” antibodies that simultaneously link to two different antigens, thereby generally bringing the different antigens in close proximity, resulting in new functionality and new therapies. it can. In general, these antibodies are made by including in the host cell a gene for each heavy and light chain. This generally results in the formation of the desired heterodimer (AB), as well as two homodimers (AA and BB (without the light chain heterodimer problem)). However, a major obstacle in bispecific antibody formation is that purifying heterodimeric antibodies from homodimeric antibodies and / or biasing heterodimer formation over homodimer formation. It is difficult to make it.

  There are many mechanisms that can be used to generate the heterodimers of the present invention. Furthermore, as will be appreciated by those skilled in the art, these mechanisms can be combined to ensure high heterodimerization. Thus, amino acid mutations that result in the production of heterodimers are termed “heterodimerization mutations”. As described below, heterodimerization mutations are steric mutations that allow purification of homodimers from heterodimers (eg, “knob and hole” or “skew” mutations described below and “charge pairs”). "Mutation") as well as "pI mutation". Useful mechanisms of heterodimerization, as incorporated herein by reference in its entirety and specifically incorporated herein for the discussion of “heterodimerization mutations” below, include “Nob and hole” (“KIH”, often referred to herein as “skew” mutations (see discussion in WO2014 / 145806), “electrostatic steering” or “charge as described in WO2014 / 145806” Pairs ”, pI mutations as described in WO2014 / 145806, and common additional Fc mutations as outlined in WO2014 / 145806 and below.

  In the present invention, there are several basic mechanisms that can facilitate the purification of heterodimeric antibodies, one of which is that AA, Rely on the use of pI mutations that allow isoelectric point purification of AB and BB dimeric proteins. Alternatively, some scaffold formats such as the “Triple F” format also allow separation based on size. It is also possible to “skew” the formation of a heterodimer rather than a homodimer, as further outlined below. Thus, combinations of steric heterodimerization mutations and pI or charge pair mutations are particularly used in the present invention.

  In general, embodiments specifically used in the present invention include skew mutations that promote heterodimer formation versus homodimer formation in combination with pI mutations that increase the pI difference between two monomers. Depends on the set of mutations.

  Furthermore, as outlined more fully below, depending on the heterodimeric antibody format, the pI mutation is contained within the monomeric constant and / or Fc domain, or is a charged linker, domain linker or scFv. Any of the linkers can be used. That is, scaffolds that utilize scFv, such as the Triple F format, can include a charged scFv linker (either positive or negative) that provides additional pI boost for purification purposes. As will be appreciated by those skilled in the art, the present invention provides pi mutations in one or both of the monomers and / or similarly charged domain linkers, but some triple F formats are only charged scFv linkers. Useful and no additional pI adjustment. Furthermore, additional amino acid manipulations for alternative functionality can also confer pI changes such as Fc, FcRn, and KO mutations.

  In the present invention that utilizes pI as a separation mechanism to allow purification of heterodimeric proteins, amino acid mutations can be introduced into one or both monomeric polypeptides, ie, monomeric One (referred to herein as “monomer A” for simplicity) can be modified to differ from monomer B, or changes in both monomers A and B can be Modifications can be made to increase the pI of body A and decrease the pI of monomer B. As discussed, pi changes in either or both monomers can be achieved by removing or adding charged residues (eg, replacing neutral amino acids with positively or negatively charged amino acid residues, eg, from glycine). Glutamic acid), by changing charged residues from positive or negative to opposite charge (eg from aspartic acid to lysine), or by changing charged residues to neutral residues (eg loss of charge, lysine to serine) Can be implemented. Some of these mutations are shown in the figure.

  Thus, this embodiment of the invention provides for causing sufficient pi changes in at least one monomer so that the heterodimer can be separated from the homodimer. As will be appreciated by those skilled in the art and as discussed further below, this is engineered to increase or decrease the “wild type” heavy chain constant region and its pI (wtA− + B or wtA−−B). Can be performed by using the mutated mutation region or by increasing one region and decreasing the other region (A + -B- or A-B +).

Thus, in general, components of some embodiments of the present invention may include at least one isoelectric point (pI), if not both, of a monomer of a dimeric protein, an amino acid substitution (a “pI mutation”). Or “pI substitution”) is an amino acid mutation in the constant region of an antibody intended to be altered by incorporation into one or both of the monomers to form a “pI antibody”. As shown herein, separation of the heterodimer from the two homodimers can be achieved when the pi of the two monomers is slightly different from 0.1 pH units, 0.2 All that differ by 0.3, 0.4, and 0.5 or more are used in the present invention.
As will be appreciated by those skilled in the art, the number of pI mutations to be included in each monomer or both monomers to obtain good separation is determined by the starting pi of the component, eg in the triple F format. Will depend in part on the scFv and the starting pI of the Fab. That is, to determine which monomer to manipulate or which “direction” (eg, more positive or more negative), the Fv sequences of the two target antigens are calculated and made from there . As is known in the art, different Fvs will have different starting pi utilized in the present invention. In general, as outlined herein, the pi is engineered to produce a total pi difference of at least about 0.1 log of each monomer, and as outlined herein, 0.2 to 0.5 is preferred. Further, as understood by those skilled in the art and outlined herein, in some embodiments, heterodimers can be separated from homodimers based on size. As shown in FIG. 2, for example, some of the formats allow for separation of heterodimers and homodimers based on size.

A. The present invention relates to heterodimer antibodies of various formats that utilize heterodimerization mutations to allow heterodimer formation and / or purification from homodimers. A heterodimeric protein comprising Some heterozygous mutations are shown in FIG.

  There are several suitable pairs of heterodimerized skew mutation sets. These mutations become “pairs” of “sets”. That is, one set of pairs is incorporated into the first monomer and the other set of pairs is incorporated into the second monomer. Of note, these sets do not necessarily behave as “knob-in-hole” mutations with a one-to-one correspondence between residues on one monomer and residues on the other monomer. That is, these set pairs promote heterodimer formation and form an interface between two monomers that prevent homodimer formation and spontaneously occur under biological conditions. 90% or more rather than the expected 50% (25% homodimer A / A: 50% heterodimer A / B: 25% homodimer B / B) To.

B. Stereomutation In some embodiments, heterodimer formation can be facilitated by the addition of steric variation. That is, changing the amino acids in each heavy chain is more likely to form a heterodimeric structure than different heavy chains associate to form a homodimer having the same Fc amino acid sequence. Appropriate steric variations are included in the drawing.

  One mechanism, commonly referred to in the art as “knob and hole”, refers to amino acid engineering that provides steric effects that favor heterodimer formation and disadvantageate homodimer formation, optionally It can also be used; this is often referred to as “knob and hall” and is described in USSN 61 / 596,846, Ridgway et al. , Protein Engineering 9 (7): 617 (1996), Atwell et al. , J .; Mol. Biol. 1997270: 26, US Pat. No. 8,216,805, all of which are hereby incorporated by reference in their entirety. These figures identify several “monomer A-monomer B” pairs that depend on “knob and hole”. In addition, Merchant et al. , Nature Biotech. 16: 677 (1998), these “knob and hole” mutations can be combined with disulfide linkages for skew formation for heterodimerization.

  Additional mechanisms used for the generation of heterodimers are described in Gunasekaran et al., Which is incorporated herein by reference in its entirety. , J .; Biol. Chem. 285 (25): 19637 (2010), often referred to as “electrostatic steering”. This is often referred to herein as a “charge pair”. In this embodiment, electrostatics is used to skew the formation towards heterodimerization. As will be appreciated by those skilled in the art, they also have the potential to affect pI, and thus purification, and therefore can be considered as pI mutations in some cases. However, they are classified as “steric mutations” because they were generated to force heterodimerization and were not used as a purification tool. These include D221E / P228E / L368E paired with D221R / P228R / K409R (eg, these are “corresponding sets of monomers”), and C220E paired with C220R / E224R / P228R / K409R. / P228E / 368E is included, but is not limited thereto.

  Additional monomer A and monomer B mutations are shown in FIG. 37, Figure and Figure and SEQ ID NO. Of US2012 / 0149876, which is specifically incorporated herein by reference, or the pI mutation outlined herein. It can be arbitrarily and independently combined with other mutations such as other steric mutations in any amount.

  In some embodiments, the steric variation outlined herein is optionally independently of any pI mutation (or other mutation, such as an Fc mutation, FcRn mutation) in one or both monomers. Can be incorporated independently and optionally included or excluded from the proteins of the invention.

  A list of suitable skew mutations can be found in FIG. 4, which shows some pairs that are particularly useful in many embodiments. Particularly useful in many embodiments are S364K / E357Q: L368D / K370S, L368D / K370S: S364K, L368E / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E A pair of sets including, but not limited to, S364K / E357Q, and T366S / L368A / Y407V: T366W (optionally bridged disulfide T366S / L368A / Y407V / Y349C: T366W / S354C). Regarding nomenclature, the pair “S364K / E357Q: L368D / K370S” means that one monomer has the double mutation set S364K / E357Q and the other has the double mutation set L368D / K370S, as described above. As such, the “twist” of these pairs depends on the starting pI.

C. Heterodimeric pI (isoelectric point) mutations As is generally understood by those skilled in the art, there are two general categories of pI mutations: those that increase the pI of a protein (basic changes) and Those that reduce the pI of proteins (acidic changes). As described herein, all combinations of these mutations can be made: one monomer can be wild-type or a mutation that does not exhibit a pi significantly different from wild-type, and the other is more It can be either basic or more acidic. Alternatively, each monomer changes one to be more basic and one to be more acidic.

  A preferred combination of pI mutations is shown in 5. As outlined herein and shown in the figure, these changes are shown relative to IgG1, but all isotypes can be changed in this way, as are isotype hybrids. If the heavy chain constant domain is derived from IgG2-4, R133E and R133Q can also be used.

In one embodiment, for example, in a bottle opener format, a preferred combination of pI mutations includes one that includes the 208D / 295E / 384D / 418E / 421D mutation (N208D / Q295E / N384D / Q418E / N421D when compared to human IgG1). It has a monomer (negative Fab side) and a second monomer (positive scFv side) containing a positively charged scFv linker containing (GKPGS) ₄ . However, as will be appreciated by those skilled in the art, the first monomer comprises a CH1 domain containing position 208. Thus, in constructs that do not include a CH1 domain (eg, in the case of a heterodimeric Fc fusion protein that does not utilize a CH1 domain for one of the domains in a dual scFv format), the preferred negative pI mutant Fc set is 295E / 384D / 418E / 421D mutation (Q295E / N384D / Q418E / N421D when compared to human IgG1).

  Thus, in some embodiments, one monomer has a set of substitutions from FIG. 5 and the other monomer is a charged linker (the monomer is a scFv or charged domain as the format indicates In order to include a linker, it has a charged scFv linker form).

1. In addition, many embodiments of the present invention rely on the “import” of a pI amino acid at a specific position from one IgG isotype to another, thus introducing unwanted immunogenicity into the variant. Reduce or eliminate the possibility of being Some of these are shown in FIG. 21 of US Patent Application Publication No. 2014/0370013, which is hereby incorporated by reference. That is, IgG1 is a common isotype of therapeutic antibodies for a variety of reasons including high effector function. However, the heavy chain constant region of IgG1 has a higher pI than that of IgG2 (8.10 vs. 7.31). By introducing an IgG2 residue at a specific position into the IgG1 backbone, the pi of the resulting monomer is reduced (or increased) and exhibits an even longer serum half-life. For example, IgG1 has glycine (pI 5.97) at position 137, IgG2 has glutamic acid (pI 3.22), and the transfer of glutamic acid affects the pI of the resulting protein. As described below, several amino acid substitutions are generally required to significantly affect the pi of the mutant antibody. However, it should be noted as discussed below that even changes in the IgG2 molecule allow for an increase in serum half-life.

  In other embodiments, to reduce the overall charge state of the resulting protein (eg, by changing from a higher pI amino acid to a lower pI amino acid), or as described in more detail below, Non-isotype amino acid changes are made to allow structural adjustments such as for stability.

  Furthermore, significant changes in each monomer of the heterodimer can be seen by pI manipulation of both heavy and light chain constant domains. As discussed herein, the pi of the two monomers differing by at least 0.5 allows for separation by ion exchange chromatography or isoelectric focusing, or other methods sensitive to isoelectric point Can be.

D. Calculate the pI The pI of each monomer can depend on the pI of the mutant heavy chain constant domain and the pI of the entire monomer including the mutant heavy chain constant domain and the fusion partner. Accordingly, in some embodiments, the change in pI is calculated based on the mutated heavy chain constant domain using the chart of FIG. 19 of US Patent Application Publication No. 2014/0370013. As discussed herein, which monomer to manipulate is generally determined by the intrinsic pI of the Fv region and the scaffold region. Alternatively, the pi of each monomer can be compared.

E. pI mutations also confer better FcRn upon in vivo ligation If pI mutations reduce monomeric pI, they can have the additional advantage of improving serum retention in vivo.

  Although still under investigation, the Fc region is thought to have a longer half-life in vivo since it is sequestered when linked to FcRn at endosomal pH 6 (Ghetie and Ward, 1997 Immunol Today. 18 ( 12): 592-598, which is incorporated by reference in its entirety). The endosomal compartment then recycles Fc to the cell surface. As the compartment opens into the extracellular space, a higher pH of about 7.4 induces the release of Fc into the blood. In mice, Dall 'Acqua et al. Showed that Fc mutations with increased FcRn ligation at pH 6 and pH 7.4 actually had reduced serum concentrations and the same half-life as wild-type Fc (Dall'Acqua et al. 2002, J. Immunol 169: 5171-5180, which is incorporated by reference in its entirety). Increased Fc affinity for FcRn at pH 7.4 is believed to prevent release of Fc into the blood. Thus, Fc mutations that increase Fc half-life in vivo ideally increase FcRn ligation at lower pH while still allowing release of Fc at higher pH. The amino acid histidine changes its charge state in the pH range of 6.0 to 7.4. It is therefore not surprising to find His residues at key positions in the Fc / FcRn complex.

  Recently, it has been suggested that antibodies with variable regions with lower isoelectric points may also have longer serum half-lives (see Igawa et al., 2010 PEDS. 23 (5): 385-392). All of which are incorporated herein by reference. However, this mechanism is still not well understood. Furthermore, the variable region varies from antibody to antibody. As described herein, constant region mutations with reduced pi and increased half-life will provide a more modular approach to improve the pharmacokinetic properties of antibodies.

F. Additional Fc mutations for additional functionality In addition to pI amino acid mutations, a variety of modifications include, but are not limited to, modification of linkage to one or more FcγR receptors, modification of linkage to FcRn receptors, etc. There are several useful Fc amino acid modifications that can be implemented for reasons.

  Accordingly, the proteins of the invention can include amino acid modifications that include heterodimerization mutations outlined herein, including pI mutations and steric mutations. Each set of mutations can be included independently or optionally excluded from a particular heterodimeric protein.

G. FcγR mutations There are several useful Fc substitutions that can be made to alter the linkage to one or more of the FcγR receptors. Substitutions that result in increased linkage and reduced linkage may be useful. For example, increased linkage to FcγRIIIa generally results in ADCC (antibody-dependent cell-mediated cytotoxicity, ie non-specific cytotoxic cells expressing FcγR recognize the linked antibody on the target cell, It is known to lead to an increase in cell-mediated reactions that cause target cell lysis. Similarly, under some circumstances, reduced linkage to FcγRIIb (inhibitory receptor) may be beneficial. Amino acid substitutions useful in the present invention include those listed in USSN 11 / 124,620 (particularly FIG. 41), USSN 11 / 174,287, USSN 11 / 396,495, USSN 11 / 538,406, all of which Are expressly incorporated herein by reference in their entirety, specifically the mutations disclosed therein. Specific mutations used include 236A, 239D, 239E, 332E, 332D, 239D / 332E, 267D, 267E, 328F, 267E / 328F, 236A / 332E, 239D / 332E / 330Y, 239D, 332E / 330L, 243A 243L, 264A, 264V, and 299T.

  Further, as specifically disclosed in USSN 12 / 341,769, the entire contents of which are incorporated herein by reference, 434S, 434A, 428L, 308F, 259I, 428L / 434S, 259I / 308F, 436I. Additional Fc substitutions used to increase linkage to FcRn receptors and increase serum half-life include, but are not limited to, / 428L, 436I or V / 434S, 436V / 428L, and 259I / 308F / 428L Exists.

H. Similarly, another category of functional mutations are “FcγR excision mutations” or “Fc knockout (FcKO or KO) mutations. In these embodiments, additional mechanisms of action are available for some therapeutic uses. It is desirable to reduce or eliminate normal linkage of the Fc domain to one or more or all Fcγ receptors (eg, FcγR1, FcγRIIa, FcγRIIb, FcγRIIIa, etc.). In particular, it is desirable to remove FCγRIIIa linkage to remove or significantly reduce ADCC activity, particularly such that one of the Fc domains contains one or more Fcγ receptor excision mutations. In the use of immunomodulating antibodies, these excision mutations are shown in FIG. L328R, E233P / L234V / L235A / G236del / S239K, E233P / L234V / L235A / G236del / S267K, E233P / L234V / L235A / G236del / S239K / A327G, E233P / L235G / S235 / In preferred embodiments using a deletion mutation selected from the group consisting of / L235A / G236del, each can be optionally included or excluded independently, wherein the deletion mutations referred to herein remove FcγR linkages. However, it should be noted that it generally does not remove FcRn linkages.

I. Combinations of Heterodimers and Fc Mutations As will be appreciated by those skilled in the art, all listed heterodimerization mutations (including skew and / or pI mutations) are all their “twist” or “single” As long as the “mer distribution” is maintained, it can be arbitrarily and independently combined in any manner. Furthermore, all of these mutations can be combined in any of the heterodimerization formats.

  In the case of pI mutations, the embodiment specifically used is shown in the drawing, but other combinations are generated according to the basic rule of changing the pI difference between the two monomers to facilitate purification. it can.

  Furthermore, any of the heterodimerization mutations, skews, and pIs can optionally be combined independently of Fc removal mutations, Fc mutations, FcRn mutations, as generally outlined herein.

VII. Useful Formats of the Invention As understood by those skilled in the art and more fully described below, as generally shown in FIG. Various configurations are possible. Some figures show a “single-ended” configuration in which one “arm” of the molecule has one type of specificity and the other “arm” has a different specificity. Other figures show a “dual-end” configuration where there is at least one type of specificity at the “top” of the molecule and one or more different specificities at the “bottom” of the molecule. Accordingly, the present invention relates to novel immunoglobulin compositions that co-link with different first and second antigens.

  As will be appreciated by those skilled in the art, the heterodimeric format of the present invention (see FIG. 2) can have different valences and is bispecific. That is, an antibody of the invention can be bivalent and bispecific (eg, heterodimers) in which the checkpoint target is linked by one ABD and the costimulatory target is linked by a second ABD. A bottle opener format), or a bispecific mAb that is a homodimer). Heterodimeric antibodies can also be trivalent and bispecific, where the first antigen is linked by two ABDs and the second antigen is linked by a second ABD (eg, , See central-scFv format and trident format). Heterodimeric antibodies can also be bispecific and tetravalent (eg, central scFv2 format and DVD-Ig format).

  Similarly, with the exception of the DVD-Ig format and the central-scFv2 format, antibodies are generally formatted so that costimulatory targets are monovalently linked.

A. Bottle Opener Format One heterodimeric scaffold specifically used in the present invention is a “Triple F” or “Bottle Opener” scaffold format. In this embodiment, one heavy chain of the antibody comprises a single chain Fv (“scFv” as defined below) and the other heavy chain in a “normal” Fab format comprising variable heavy and light chains. is there. This structure is often referred to herein as the “triple F” format (scFv-Fab-Fc) or “bottle opener” format because of its visual approximate similarity to the bottle opener. As described more fully below, the two chains are brought together by the use of amino acid mutations in constant regions that facilitate the formation of heterodimeric antibodies (eg, Fc domain, CH1 domain and / or hinge region). .

  There are several distinct advantages to the current “Triple F” format. As is known in the art, antibody analogs that rely on two scFv constructs often have stability and aggregation problems, which in the present invention is the “normal” heavy and light chain pairing. It can be reduced by adding. In addition, there is a problem of mispairing heavy chain and light chain (eg heavy chain 1 paired with light chain 2) as opposed to a format that relies on two heavy chains and two light chains. do not do.

  Many of the embodiments outlined herein generally include variable heavy and variable light chain domains covalently linked using scFv linkers (many charged, but not all). The scFv is covalently linked to the first Fc domain, usually via a domain linker (this is outlined herein). As such, it may be uncharged or charged, exogenous or endogenous (eg, all or part of a natural hinge domain) The second monomer in the bottle opener format is a heavy chain Yes, the composition further comprises a light chain.

  In addition, the bottle opener format Fc domain generally contains skew mutations (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S. : S364K, L368E / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E357L, K370S: S364K / E357Q, T366S / L368A / Y367V: T366S / T366 / 36 Selected from the group consisting of S354C, optionally including a deletion mutation (including that shown in FIG. 6), optionally including a charged scFv linker (including that shown in FIG. 8), Includes pI mutations (including those shown in FIG. 5).

  In some embodiments, the bottle opener format includes a skew mutation, a pi mutation, and a removal mutation. Thus, some embodiments include a bottle opener format that includes: a) a charged scFv linker (preferred in some embodiments, having the + H sequence of FIG. 8), skew mutation S364K / E357Q, removal mutation E233P / L234V / L235A / G236del / S267K, and a first monomer (“scFv monomer”) that contains Fv linked to one target outlined herein, b) skew mutation L368D / K370S, pI Along with the mutation N208D / Q295E / N384D / Q418E / N421D, the removal mutation E233P / L234V / L235A / G236del / S267K, and a variable light chain domain that creates an Fv that links to the second target outlined herein A second monomer comprising a heavy chain domain "Fab monomer"), c) the light chain. In this particular embodiment, suitable monomeric Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA. ICOS x TIGIT, TIGIT x ICOS, PD-1 x ICOS, PD-L1 x ICOS, CTLA-4 x ICOS, LAG-3 x ICOS, TIM-3 x ICOS, BTLA x ICOS, OX40 x TIGIT, OX40 x PD -1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BTLA, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG-3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, GITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA -4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4-1BB, CTLA-4 × 4 -1BB, LAG-3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB are included (Fab listed first, scFv listed second).

  Thus, some embodiments include a bottle opener format that includes: a) a charged scFv linker (preferred in some embodiments, having the + H sequence of FIG. 8), skew mutation S364K / E357Q, removal mutation E233P / L234V / L235A / G236del / S267K, and a first monomer (“scFv monomer”) that contains Fv linked to one target outlined herein, b) skew mutation L368D / K370S, pI Along with the mutation N208D / Q295E / N384D / Q418E / N421D, the removal mutation E233P / L234V / L235A / G236del / S267K, and a variable light chain domain that creates an Fv that links to the second target outlined herein A second monomer comprising a heavy chain domain "Fab monomer"), c) the light chain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the bottle opener format includes a skew mutation, a pi mutation, a removal mutation, and an FcRn mutation. Thus, some embodiments include a bottle opener format that includes: a) a charged scFv linker (preferred in some embodiments, having the + H sequence of FIG. 8), skew mutation S364K / E357Q, removal mutation E233P / L234V / L235A / G236del / S267K, FcRn mutation M428L / N434S, and a first comprising an Fv linked to a first receptor outlined herein (either a costimulatory receptor or a checkpoint receptor) Monomer ("scFv monomer"), b) skew mutation L368D / K370S, pI mutation N208D / Q295E / N384D / Q418E / N421D, removal mutation E233P / L234V / L235A / G236del / S267K, FcRn mutation M428L / N43 S and, together with the variable light chain domain, includes a variable heavy chain domain that creates an Fv linked to the second receptor outlined herein (the other of the costimulatory receptor or the checkpoint receptor) A second monomer (“Fab monomer”), c) a light chain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  Specifically, FIG. 9 shows several bottle opener “backbone” sequences that lack Fv sequences that can be used in the present invention. That is, the Fv sequences of the scFv part and the Fab part are ICOS and PD-1, ICOS and CTLA-4, ICOS and LAG-3, ICOS and TIM-3, ICOS and PD-L1, ICOS and BTLA, ICOS and TIGIT, GITR and TIGIT, GITR and PD-1, GITR and CTLA-4, GITR and LAG-3, GITR and TIM-3, GITR and PD-L1, GITR and BTLA, OX40 and PD-1, OX40 and TIGIT, OX40 CTLA-4, IC OX40 OS and LAG-3, OX40 and TIM-3, OX40 and PD-L1, OX40 and BTLA, 4-1BB and PD-1, 4-1BB and CTLA-4, 4-1BB and LAG- 3, 4-1BB and TIM-3, 4-1BB and PD-L1, TIGIT and 4-1BB, and And any combination of 4-1BB and BTLA can be used, combined with any or all of skeletons 1 to 10, and skeleton 1 is particularly used in these combinations.

  With respect to the bottle opener scaffold 1 from FIG. 9 (optionally including the 428L / 434S mutation), specific Fv combinations for use in the present invention include ICOS and PD-1, ICOS and PD-L1, and ICOS CTLA-4 is included.

  For the bottle opener scaffold 1 from FIG. 9 (optionally including the 428L / 434S mutation), specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and FIGS. 20, those shown in FIGS. 24, 68, and 77, and those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665. It is not limited.

  For the bottle opener scaffold 1 from FIG. 9 (optionally including the 428L / 434S mutation), specific ABDs linked to human GITR include those in FIGS. 18, 72, and 73, and SEQ ID NO: 26282. -26290, but are not limited to these.

  For the bottle opener backbone 1 from FIG. 9 (optionally including the 428L / 434S mutation), specific ABDs linked to OX40 are listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281. However, it is not limited to these.

  For the bottle opener scaffold 1 from FIG. 9 (optionally including the 428L / 434S mutation), specific ABDs linked to human 4-1BB include FIGS. 16, 72, and 73, and SEQ ID NO: 26262. 2671, but is not limited thereto.

  With respect to the bottle opener backbone 1 from FIG. 9 (optionally including the 428L / 434S mutation), specific ABDs linked to human PD-L1 include FIGS. 15, 73 and 78, and SEQ ID NO: 3961- 4432, but is not limited to.

  Specific bottle opener embodiments are outlined below.

B. mAb-Fv format One heterodimeric scaffold particularly used in the present invention is the mAb-Fv format. In this embodiment, the format involves the use of a C-terminal linkage of an “additional” variable heavy chain domain to one monomer and a C-terminal linkage of an “additional” variable light chain domain to the other monomer. Dependent, thus forming a third antigen linking domain (ie, an “additional” Fv domain), the two monomeric Fab portions linking to one checkpoint target, and the “additional” Fv domain co-stimulating Link to the target.

  In this embodiment, the first monomer comprises a first heavy chain comprising a first constant heavy chain domain comprising a first variable heavy chain domain and a first Fc domain, wherein the first variable light chain The domain is covalently linked to the C-terminus of the first Fc domain using a domain linker (vh1-CH1-hinge-CH2-CH3- [optional linker] -vl2). The second monomer is covalently linked to the C-terminus of the second Fc domain using a second variable heavy chain domain, a second constant heavy chain domain comprising a second Fc domain, and a domain linker. Contains three variable heavy chain domains (vh1-CH1-hinge-CH2-CH3- [optional linker] -vh2). This embodiment further utilizes a common light chain that includes a variable light chain domain and a constant light chain domain, which associate with the heavy chain to form two identical Fabs comprising two identical Fvs. The two C-terminal linking variable domains constitute an “additional” third Fv. For many embodiments herein, these constructs include skew mutations, pI mutations, elimination mutations, additional Fc mutations, etc., as desired and described herein. In this embodiment, suitable Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA, ICOS × TIGIT, TIGIT × ICOS, PD-1 × ICOS, PD-L1 × ICOS, CTLA-4 × ICOS, LAG-3 × ICOS, TIM-3 × ICOS, BTLA × ICOS, OX40 × TIGIT, OX40 × PD-1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BTLA, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG- 3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, ITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA-4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4, 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4-1BB, CTLA-4 × 4-1BB, LAG- 3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB are included (Fab listed first, "additional" Fv listed second).

  The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  In addition, the Fb domain in mAb-Fv format contains a skew mutation (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S. : S364K, L368E / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E357L, K370S: S364K / E357Q, T366S / L368A / Y367V: T366S / T366 / 36 Optionally selected from the group consisting of S354C, including a deletion mutation (error including those shown in FIG. 6! No referrer found), and optionally a charged scFv linker (shown in FIG. 8) Includes including) from, comprising a heavy chain is pI mutations (errors, including those shown in FIG. 5! No reference source is found.).

  In some embodiments, the mAb-Fv format includes a skew mutation, a pi mutation, and a removal mutation. Accordingly, some embodiments comprise a mAb-Fv format comprising: a) a skew variable S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, wherein the first variable light chain of the light chain A first monomer comprising a first variable heavy chain domain and a second variable heavy chain domain that together with the domain comprise an Fv linked to a first checkpoint inhibitor, b) skew mutation L368D / K370S, a first receptor as outlined herein, comprising a pI mutation N208D / Q295E / N384D / Q418E / N421D, a deletion mutation E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain ( Either costimulatory receptors or checkpoint receptors) A first variable heavy chain domain that constitutes an Fv to be coupled, and a second variable heavy chain that together with the second variable heavy chain constitutes an Fv (ABD) linked to a second receptor (the other of a costimulatory or checkpoint receptor) A second monomer comprising a variable light chain, and c) a light chain comprising a first variable light chain domain and a constant light chain domain. Particularly useful in some embodiments of this format are ICOS × PD-1, PD-1 × ICOS, PD-L1 × ICOS, ICOS × PD-L1, GITR × PD (in the order of Fab-scFv). -1, OX40 × PD-1 and 4-1BB × PD-1. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the mAb-Fv format comprises a skew mutation, a pi mutation, a removal mutation, and an FcRn mutation. Accordingly, some embodiments comprise a mAb-Fv format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, and a light chain A first monomer comprising a first variable heavy chain domain and a first variable heavy chain domain comprising an Fv linked to a first checkpoint inhibitor together with a first variable light chain domain, b) Including skew mutation L368D / K370S, pI mutation N208D / Q295E / N384D / Q418E / N421D, removal mutation E233P / L234V / L235A / G236del / S267K, FcRn mutation M428L / N434S, together with the first variable light chain domain As outlined in A first variable heavy chain domain constituting an Fv linked to a first checkpoint inhibitor, and a second variable heavy chain constituting an Fv (ABD) linked to a second checkpoint inhibitor. A second monomer comprising a variable light chain, and c) a light chain comprising a first variable light chain domain and a constant light chain domain. Particularly useful in some embodiments of this format are ICOS × PD-1, ICOS × PD-L1, GITR × PD-1, OX40 × PD-1 and 4-1BB (in the order of Fab-scFv). × PD-1.

  For mAb-Fv sequences similar to mAb-scFv backbone 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human ICOS are [ICOS] _H0L0 and [ICOS] _H0.66_L0, and 19, 20, 24, 68, and 77, and those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  With respect to Fb sequences of mAb-similar to mAb-scFv scaffold 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human PD-L1 are shown in FIGS. 15, 73, and 78. As well as those shown in SEQ ID NOs: 3961-4432.

  Specific ABDs linked to human GITR with respect to Fb sequences of mAb-similar to mAb-scFv backbone 1 from FIG. 75 (optionally including M428L / N434S) are those in FIGS. 18, 72, and 73. As well as those listed in SEQ ID NOs: 26282 to 26290.

  With respect to the Fb sequence of mAb-similar to mAb-scFv backbone 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human 4-1BB are shown in FIGS. 16, 72, and 73. As well as those in SEQ ID NOs: 26262 to 2671.

  Specific ABDs linked to human OX40 for mAb-Fv sequences similar to mAb-scFv backbone 1 from FIG. 75 (optionally including M428L / N434S) are those in FIGS. 17, 72, and 73. As well as those listed in SEQ ID NOs: 26272-26281.

  With respect to Fb sequences of mAb-similar to mAb-scFv scaffold 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human PD-L1 are from FIGS. 15, 73, and 78. As well as those from SEQ ID NOs: 3961-4432.

C. mAb-scFv
One heterodimeric scaffold specifically used in the present invention is the mAb-scFv format. In this embodiment, the format relies on the use of a scFv C-terminal linkage to one of the monomers, thus forming a third antigen-binding domain, where the Fab portions of the two monomers are Linked to one receptor target, the “additional” scFv domain is linked to the other receptor target (generally a monovalently linked costimulatory receptor).

  In this embodiment, the first monomer is a first heavy chain (variable heavy) having a C-terminal covalently linked scFv comprising an scFv variable light chain domain, an scFv linker and an scFv variable heavy chain domain in either orientation. (Including chain domain and constant domain) (vh1-CH1-hinge-CH2-CH3- [optional linker] -vh2-scFv linker-vl2 or vh1-CH1-hinge-CH2-CH3- [optional linker]- vl2-scFv linker-vh2). This embodiment further utilizes a common light chain comprising a variable light chain domain and a constant light chain domain, which associates with the heavy chain to form two identical Fabs linked to one of the target receptors. . For many embodiments herein, these constructs include skew mutations, pI mutations, elimination mutations, additional Fc mutations, etc., as desired and described herein. In this embodiment, suitable Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA, ICOS × TIGIT, TIGIT × ICOS, PD-1 × ICOS, PD-L1 × ICOS, CTLA-4 × ICOS, LAG-3 × ICOS, TIM-3 × ICOS, BTLA × ICOS, OX40 × TIGIT, OX40 × PD-1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BTLA, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG- 3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, ITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA-4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4, 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4-1BB, CTLA-4 × 4-1BB, LAG- 3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB are included (Fab listed first, scFv listed second).

  The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  In addition, the Fb domain in mAb-scFv format contains a skew mutation (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S. : S364K, L368E / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E357L, K370S: S364K / E357Q, T366S / L368A / Y367V: T366S / T366 / 36 Selected from the group consisting of S354C, optionally including a deletion mutation (including that shown in FIG. 6), optionally including a charged scFv linker (including that shown in FIG. 8), wherein the heavy chain is p Mutations, including the (error including those shown in FIG. 5! Reference source not found.).

  In some embodiments, the mAb-scFv format includes a skew mutation, a pi mutation, and a removal mutation. Accordingly, some embodiments include a bottle opener format comprising: a) a skew variable S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, and the first variable light chain domain of the light chain And a first monomer comprising a first variable heavy chain domain constituting an Fv linked to a first receptor and an scFv linked to a second receptor, b) a skew mutation L368D / K370S, pI The mutation N208D / Q295E / N384D / Q418E / N421D, including the removal mutation E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain, linked to a first receptor as outlined herein A second monomer comprising a first variable heavy chain domain comprising an Fv , And c) light chain comprising a first variable light chain domain and the constant light chain domain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the mAb-scFv format comprises a skew mutation, a pi mutation, a removal mutation, and an FcRn mutation. Thus, some embodiments comprise a mAb-scFv format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, A first monomer comprising, together with a first variable light chain domain, a first variable heavy chain domain constituting an Fv linked to a first receptor and an scFv linked to a second receptor, b) Including skew mutation L368D / K370S, pI mutation N208D / Q295E / N384D / Q418E / N421D, removal mutation E233P / L234V / L235A / G236del / S267K, FcRn mutation M428L / N434S, together with the first variable light chain domain No. as outlined in First second monomer comprising the variable heavy chain domain, and c) light chain comprising a first variable light chain domain and the constant light chain domain constituting the Fv that couples to the checkpoint inhibitors.

  In the mAb-scFv scaffold 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human ICOS are those shown in FIG. 19, FIG. 20, FIG. 24, FIG. 68, and FIG. And those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In the mAb-scFv backbone 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human GITR are shown in FIG. 18, FIG. 72, and FIG. 73, as well as SEQ ID NOs: 26282-26290. Listed.

  In the mAb-scFv scaffold 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human 4-1BB include those in FIGS. 16, 72, and 73, and SEQ ID NO: 26262 Including those in 2671.

  In the mAb-scFv backbone 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human OX-40 include those in FIGS. 17, 72, and 73, and SEQ ID NO: 26272 26281.

  In the mAb-scFv scaffold 1 from FIG. 75 (optionally including M428L / N434S), specific ABDs linked to human PD-L1 include FIGS. 15, 73 and 78, and SEQ ID NOs: 3961-4432 .

D. Central scFv
One heterodimeric scaffold specifically used in the present invention is the central-scFv format. In this embodiment, the format relies on the use of an inserted scFv domain, thus forming a third antigen-binding domain, where the two monomeric Fab portions are linked to one receptor target. , The “additional” scFv domain is linked to another (also typically a monovalently linked costimulatory receptor). The scFv domain is inserted between the Fc domain and the CH1-Fv region of one of the monomers, thus providing a third antigen linking domain.

  In this embodiment, one monomer is a first variable heavy chain domain, a CH1 domain (and any hinge) and an Fc domain, together with a scFv comprising a scFv variable light chain domain, a scFv linker and a scFv variable heavy chain domain. A first heavy chain comprising The scFv is covalently linked between the C-terminus of the CH1 domain of the heavy chain constant domain and the N-terminus of the first Fc domain using an optional domain linker (vh1-CH1- [optional linker] -Vh2-scFv linker-vl2- [any linker including hinge] -CH2-CH3, or the opposite orientation to scFv, vh1-CH1- [optional linker] -vl2-scFv linker-vh2- [including hinge Optional linker] -CH2-CH3). In some embodiments, the optional linker is a hinge or fragment thereof. The other monomer is the standard Fab side (eg, vh1-CH1-hinge-CH2-CH3). This embodiment further utilizes a common light chain comprising a variable light chain domain and a constant light chain domain, which associate with the heavy chain to form two identical Fabs linked to a checkpoint inhibitor. For many embodiments herein, these constructs include skew mutations, pI mutations, elimination mutations, additional Fc mutations, etc., as desired and described herein. In this embodiment, suitable Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA, ICOS × TIGIT, TIGIT × ICOS, PD-1 × ICOS, PD-L1 × ICOS, CTLA-4 × ICOS, LAG-3 × ICOS, TIM-3 × ICOS, BTLA × ICOS, OX40 × TIGIT, OX40 × PD-1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BTLA, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG- 3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, ITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA-4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4, 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4-1BB, CTLA-4 × 4-1BB, LAG- 3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB are included (Fab listed first, scFv listed second).

  The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  In addition, the Fc domain of the central scFv format generally contains a skew mutation (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S. : S364K, L368E / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E357L, K370S: S364K / E357Q, T366S / L368A / Y367V: T366S / T366 / 36 Selected from the group consisting of S354C, optionally including a deletion mutation (including that shown in FIG. 6), optionally including a charged scFv linker (including that shown in FIG. 8), wherein the heavy chain is Including I mutations (including those shown in FIG. 5).

  In some embodiments, the central scFv format includes a skew mutation, a pi mutation, and a removal mutation. Thus, some embodiments comprise a central scFv format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, and the first variable light chain domain of the light chain And a first monomer comprising a first variable heavy chain domain constituting an Fv linked to a first receptor, b) a skew mutation L368D / K370S, a pI mutation N208D / Q295E / N384D / Q418E / N421D, A first variable heavy chain comprising an excision mutation E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain, comprising an Fv linked to a second receptor as outlined herein A second monomer comprising a domain, and c) a first variable light chain domain and Light chain comprising the constant light chain domain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the central scFv format includes skew mutations, pI mutations, removal mutations, and FcRn mutations. Accordingly, some embodiments include a central-scFv format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, A first monomer comprising, together with a first variable light chain domain, a first variable heavy chain domain constituting an Fv linked to a first receptor, and an scFv domain linked to a second receptor; b ) Including skew mutation L368D / K370S, pI mutation N208D / Q295E / N384D / Q418E / N421D, removal mutation E233P / L234V / L235A / G236del / S267K, FcRn mutation M428L / N434S, together with the first variable light chain domain, the present specification Will be outlined in the book A second monomer comprising a first variable heavy chain domain comprising an Fv linked to a first checkpoint inhibitor, and c) a light chain comprising a first variable light chain domain and a constant light chain domain . In this embodiment, suitable Fv pairs include ICOS × PD-1, PD-1 × ICOS, ICOS × PD-L1, PD-L1 × ICOS, ICOS × CTLA-4, and CTLA-4 × ICOS ( Fab is listed first and scFv is listed second).

  For the central-scFv sequence using the central-scFv sequence of FIG. 55, suitable Fvs linked to ICOS are those shown in FIGS. 19, 20, 24, 68, and 77, and SEQ ID NOs: 27869- Including but not limited to those shown in 28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  With respect to the central-scFv sequence using the central-scFv sequence of FIG. 55, suitable Fvs that link to PD-L1 are shown in FIGS. 15, 73, and 78, and are shown in SEQ ID NOs: 3961-4432 Including, but not limited to.

  For the central-scFv sequence using the central-scFv sequence of FIG. 55, suitable Fvs that link to GITR are those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282-26290. Including but not limited to.

  For the central-scFv sequence using the central-scFv sequence of FIG. 55, suitable Fvs that link to OX40 include those listed in FIGS. 17, 72, and 73, and SEQ ID NOs: 26272-26281. It is not limited to.

  For a central-scFv sequence using the central-scFv sequence of FIG. 55, suitable Fvs that link to 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671. .

E. Middle-scFv2
One heterodimeric scaffold specifically used in the present invention is the central-scFv2 format, which is bispecific and trispecific. In this embodiment, the format relies on the use of two inserted scFv domains, thus forming third and fourth antigen-binding domains, where the two monomeric Fab portions are one receptor Linked to the body target, the “additional” scFv domain is linked to another. The scFv domain is inserted between the monomeric Fc domain and the CH1-Fv region.

  In this embodiment, both monomers have a first variable heavy chain domain, a CH1 domain (and optional hinge) and an Fc domain, together with an scFv comprising an scFv variable light chain domain, an scFv linker and an scFv variable heavy chain domain. A first heavy chain comprising The scFv is covalently linked between the C-terminus of the CH1 domain of the heavy chain constant domain and the N-terminus of the first Fc domain using an optional domain linker (vh1-CH1- [optional linker] -Vh2-scFv linker-vl2- [any linker including hinge] -CH2-CH3, or the opposite orientation to scFv, vh1-CH1- [optional linker] -vl2-scFv linker-vh2- [including hinge Optional linker] -CH2-CH3). In some embodiments, the optional linker is a hinge or fragment thereof. This embodiment further utilizes a common light chain comprising a variable light chain domain and a constant light chain domain, which associate with the heavy chain to form two identical Fabs that are linked to the receptor. For many embodiments herein, these constructs include skew mutations, pI mutations, elimination mutations, additional Fc mutations, etc., as desired and described herein. In this embodiment, suitable Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA, ICOS × TIGIT, TIGIT × ICOS, PD-1 × ICOS, PD-L1 × ICOS, CTLA-4 × ICOS, LAG-3 × ICOS, TIM-3 × ICOS, BTLA × ICOS, OX40 × TIGIT, OX40 × PD-1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BTLA, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG- 3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, ITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA-4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4, 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4-1BB, CTLA-4 × 4-1BB, LAG- 3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB are included (Fab listed first, scFv listed second).

  The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  In addition, the Fc domain of the central scFv2 format generally contains a skew mutation (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S. : S364K, L368E / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E357L, K370S: S364K / E357Q, T366S / L368A / Y367V: T366S / T366 / 36 A heavy chain selected from the group consisting of S354C, optionally including a deletion mutation (including that shown in FIG. 6), optionally including a charged scFv linker (including that shown in FIG. 8); pI containing mutations (including those shown in FIG. 5).

  In some embodiments, the central scFv2 format includes a skew mutation, a pi mutation, and a removal mutation. Thus, some embodiments comprise a central scFv format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, and the first variable light chain domain of the light chain And a first monomer comprising a first variable heavy chain domain constituting an Fv linked to a first receptor, b) a skew mutation L368D / K370S, a pI mutation N208D / Q295E / N384D / Q418E / N421D, A first variable heavy chain comprising an excision mutation E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain, comprising an Fv linked to a second receptor as outlined herein A second monomer comprising a domain, and c) a first variable light chain domain and Light chain comprising the constant light chain domain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the central scFv2 format includes a skew mutation, a pi mutation, a removal mutation, and an FcRn mutation. Accordingly, some embodiments include a central-scFv format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, A first monomer comprising, together with a first variable light chain domain, a first variable heavy chain domain constituting an Fv linked to a first receptor, and an scFv domain linked to a second receptor; b ) Including skew mutation L368D / K370S, pI mutation N208D / Q295E / N384D / Q418E / N421D, removal mutation E233P / L234V / L235A / G236del / S267K, FcRn mutation M428L / N434S, together with the first variable light chain domain, the present specification Will be outlined in the book A second monomer comprising a first variable heavy chain domain comprising an Fv linked to a first checkpoint inhibitor, and c) a light chain comprising a first variable light chain domain and a constant light chain domain . In this embodiment, suitable Fv pairs include ICOS × PD-1, PD-1 × ICOS, ICOS × PD-L1, PD-L1 × ICOS, ICOS × CTLA-4, and CTLA-4 × ICOS ( Fab is listed first, scFv is listed second).

  For the central-scFv2 sequence using the central-scFv2 sequence of FIG. 55, suitable Fvs linked to ICOS are those shown in FIGS. 19, 20, 24, 68, and 77, and SEQ ID NOs: 27869- Including but not limited to those shown in 28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  For the central-scFv2 sequence using the central-scFv2 sequence of FIG. 55, suitable Fvs that link to PD-L1 are shown in FIGS. 15, 73, and 78, and are shown in SEQ ID NOs: 3961-4432 Including, but not limited to.

  For the central-scFv2 sequence using the central-scFv sequence of FIG. 55, suitable Fvs that link to GITR are those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282-26290. Including but not limited to.

  For the central-scFv2 sequence using the central-scFv sequence of FIG. 55, suitable Fvs that link to OX40 include those listed in FIGS. 17, 72, and 73 and SEQ ID NOs: 26272-26281. It is not limited to.

  For a central-scFv2 sequence using the central-scFv sequence of FIG. 55, suitable Fvs that link to 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671. .

F. 1 arm mAb
As mentioned above, surprisingly and unexpectedly, monovalent costimulatory antibodies comprising a single ABD against the target show efficacy in activating T cells.

  Accordingly, in some embodiments, the present invention provides monovalent monospecific antibodies, as shown in FIG. 2N of the figure, that include (for stability) heterodimeric Fc domains. In this embodiment, in this embodiment, one monomer contains only the Fc domain and the other monomer is HC (VH1-CH1-hinge-CH2-CH3). This embodiment further utilizes a light chain comprising a variable light chain domain and a constant light chain domain, which associate with the heavy chain to form a Fab. For many embodiments herein, these constructs include skew mutations, pI mutations, elimination mutations, additional Fc mutations, etc., as desired and described herein. In this embodiment, a suitable ABD is linked to a costimulatory receptor such as ICOS, GITR, OX40, or 4-1BB.

  The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  In addition, Fc domains generally contain skew mutations (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S: S364K, L368E. / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E357L, K370S: S364K / E357Q, T366S / L368A / Y407V: T366W / T366S / L368V / 35 Optionally including a deletion mutation (including that shown in FIG. 6) and the heavy chain including a pI mutation (including that shown in FIG. 5).

  In some embodiments, the one-arm scFv-mAb format comprises a skew mutation, a pI mutation, and a removal mutation. Accordingly, some embodiments include a format comprising: a) a first (Fc) monomer comprising a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, b) a skew. Costimulatory receptor as outlined herein, including mutation L368D / K370S, pI mutation Q295E / N384D / Q418E / N421D, excision mutation E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain A second monomer comprising a first variable heavy chain domain comprising an Fv linked to the body.

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

  The specific “one arm mAb” is shown in the drawings and sequence listing.

G. Bispecific mAb
One heterodimeric scaffold specifically used in the present invention is a bispecific mAb format that is bispecific and trispecific. In this embodiment, the format relies on the production of separate homodimeric antibodies that are subsequently recombined. In this format, one HC-LC pair (VH1-CH1-hinge-CH2-CH3 and VL1-LC) and a second Hc-LC pair (VH2-CH1-hinge-CH2-CH3 and VL2-LC), for example There are two different heavy chains and two different light chains. Reference is made to Example 5I (d).

  In this format, the appropriate pairs are ICOS and PD-1, ICOS and CTLA-4, ICOS and LAG-3, ICOS and TIM-3, ICOS and PD-L1, ICOS and BTLA, ICOS and TIGIT, GITR and TIGIT. GITR and PD-1, GITR and CTLA-4, GITR and LAG-3, GITR and TIM-3, GITR and PD-L1, GITR and BTLA, OX40 and PD-1, OX40 and TIGIT, OX40 and CTLA-4 IC OX40 OS and LAG-3, OX40 and TIM-3, OX40 and PD-L1, OX40 and BTLA, 4-1BB and PD-1, 4-1BB and CTLA-4, 4-1BB and LAG-3, 4 -1BB and TIM-3, 4-1BB and PD-L1, TIGIT and 4-1BB, and 4-1BB and BT Includes LA.

  In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes an ICOS ABD of H0.66_L0 combined with ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

H. Central-Fv format One heterodimeric scaffold that is particularly used in the present invention is the central-Fv format shown in FIG. In this embodiment, the format relies on the use of an inserted Fv domain, thus forming an “additional” third antigen-binding domain, where the two monomeric Fab portions are one receptor. And the “additional” central-Fv domain is linked to another (generally a costimulatory receptor). The Fv domain is inserted between the monomeric Fc domain and the CH1-Fv region, thus providing a third antigen-binding domain, wherein each monomer comprises a component of Fv (eg, one ) Monomers contain a variable heavy chain domain and the other contains a variable light chain of an “additional” central Fv domain.

  In this embodiment, one monomer comprises a first variable heavy chain domain, a first heavy chain comprising a CH1 domain, and an Fc domain, and an additional variable light chain domain. The additional variable light chain domain is covalently linked between the C-terminus of the CH1 domain of the heavy chain constant domain and the N-terminus of the first Fc domain using a domain linker (vh1-CH1- [optional linker ] -Vl2-hinge-CH2-CH3). Other monomers include a first variable heavy chain domain, a first heavy chain comprising a CH1 domain, and an Fc domain, and an additional variable heavy chain domain (vh1-CH1- [optional linker] -vh2- Hinge-CH2-CH3). The domain of the additional variable heavy chain domain is covalently linked between the C-terminus of the CH1 domain of the heavy chain constant domain and the N-terminus of the first Fc domain using a domain linker. This embodiment utilizes a common light chain comprising a variable light chain domain and a constant light chain domain, which associates with the heavy chain to form two identical Fabs each linked to a receptor. The additional variable heavy chain domain and the additional variable light chain domain form an “additional” central Fv that is linked to a second receptor. For many embodiments herein, these constructs include skew mutations, pI mutations, elimination mutations, additional Fc mutations, etc., as desired and described herein. In this embodiment, suitable Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA, ICOS × TIGIT, TIGIT × ICOS, PD-1 × ICOS, PD-L1 × ICOS, CTLA-4 × ICOS, LAG-3 × ICOS, TIM-3 × ICOS, BTLA × ICOS, OX40 × TIGIT, OX40 × PD-1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BTLA, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG- 3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, ITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA-4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4, 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4-1BB, CTLA-4 × 4-1BB, LAG- 3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB are included (Fab listed first, "additional" central Fv listed second).

  The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  In addition, the Fc domain of the central-Fv format generally contains skew mutations (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S: S364K; L368E / K370S: S364K; T411T / E360E / Q362E: D401K; L368D / K370S: S364K / E357L; K370S: S364K / E357Q; / S354C, optionally including a deletion mutation (including that shown in FIG. 6), optionally including a charged scFv linker (including that shown in FIG. 8), and the heavy chain is p Mutations (including those shown in FIG. 5).

  In some embodiments, the central-Fv format includes a skew mutation, a pi mutation, and a removal mutation. Thus, some embodiments comprise a central scFv format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, and the first variable light chain domain of the light chain And a first monomer comprising a first variable heavy chain domain constituting an Fv linked to a first receptor, b) a skew mutation L368D / K370S, a pI mutation N208D / Q295E / N384D / Q418E / N421D, A first variable heavy chain comprising an excision mutation E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain, comprising an Fv linked to a second receptor as outlined herein A second monomer comprising a domain, and c) a first variable light chain domain and Light chain comprising the constant light chain domain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the central-Fv format includes skew mutations, pi mutations, removal mutations, and FcRn mutations. Accordingly, some embodiments comprise a central-scFv format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, A first monomer comprising, together with a first variable light chain domain, a first variable heavy chain domain constituting an Fv linked to a first receptor, and an scFv domain linked to a second receptor; b ) Including skew mutation L368D / K370S, pI mutation N208D / Q295E / N384D / Q418E / N421D, removal mutation E233P / L234V / L235A / G236del / S267K, FcRn mutation M428L / N434S, together with the first variable light chain domain, the present specification Will be outlined in the book A second monomer comprising a first variable heavy chain domain comprising an Fv linked to a first checkpoint inhibitor, and c) a light chain comprising a first variable light chain domain and a constant light chain domain . In this embodiment, suitable Fv pairs include ICOS × PD-1, PD-1 × ICOS, ICOS × PD-L1, PD-L1 × ICOS, ICOS × CTLA-4, and CTLA-4 × ICOS ( Fab is listed first and scFv is listed second).

  For the central-Fv format, suitable Fvs linked to ICOS are those shown in FIGS. 19, 20, 24, 68, and 77, and SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, Including but not limited to those shown in 28549-28556, and 28557-28665.

  For the central-Fv format, suitable Fvs linked to PD-L1 include, but are not limited to, those shown in FIGS. 15, 73, and 78, and those shown in SEQ ID NOs: 3961-4432.

  For the central-Fv format, suitable Fvs that link to GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282-26290.

  With respect to the central-Fv format, suitable Fvs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72, and 73, and SEQ ID NOs: 26272-26281.

  For the central-Fv format, suitable Fvs that link to 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  For the central-FV format, suitable Fvs linked to PD-L1 include, but are not limited to, those of FIGS. 15, 73 and 78, and those of SEQ ID NOs: 3961-4432.

I. 1 arm center-scFv
One heterodimeric scaffold that is specifically used in the present invention is the one-arm, central-scFv format shown in FIG. 1C. In this embodiment, one monomer contains only the Fc domain, and the other monomer contains the Fab domain (first antigen linking domain), scFv domain (second antigen linking domain), and Fc domain. Here, the scFv domain is inserted between the Fc domain and the Fc domain. In this format, the Fab moiety is linked to one receptor target and the scFv is linked to another.

  In this embodiment, one monomer comprises a first variable heavy chain domain, a CH1 domain and an Fc domain together with a scFv comprising a scFv variable light chain domain, a scFv linker and a scFv variable heavy chain domain. Includes chains. scFv is VH1-CH1- [arbitrary domain linker] -VH2-scFv linker-VL2- [arbitrary domain linker] -CH2-CH3 or VH1-CH1- [arbitrary domain linker] -VL2-scFv linker- VH2- [Arbitrary domain linker] -Shared between the C-terminus of the CH1 domain of the heavy chain constant domain and the N-terminus of the first Fc domain using a domain linker in either orientation of CH2-CH3 It is connected. The second monomer contains the Fc domain (CH2-CH3). This embodiment further utilizes a light chain comprising a variable light chain domain and a constant light chain domain, which associate with the heavy chain to form a Fab. For many embodiments herein, these constructs include skew mutations, pI mutations, elimination mutations, additional Fc mutations, etc., as desired and described herein. In this embodiment, suitable Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA, ICOS × TIGIT, TIGIT × ICOS, PD-1 × ICOS, PD-L1 × ICOS, CTLA-4 × ICOS, LAG-3 × ICOS, TIM-3 × ICOS, BTLA × ICOS, OX40 × TIGIT, OX40 × PD-1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BTLA, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG- 3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, ITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA-4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4, 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4-1BB, CTLA-4 × 4-1BB, LAG- 3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB are included (Fab listed first, scFv listed second).

  The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  In addition, the one-arm central-scFv format Fc domain generally contains a skew mutation (eg, a set of amino acid substitutions as shown in FIG. 4), and a particularly useful skew mutation is L368D / K370S: S364K; L368E / K370S: S364K; T411T / E360E / Q362E: D401K; L368D / K370S: S364K / E357L; K370S: S364K / E357Q; Y349C: selected from the group consisting of T366W / S354C, optionally including a removal mutation (including that shown in FIG. 6), and optionally a charged scFv linker (including that shown in FIG. 8) Look, comprising a heavy chain pI mutations (including those shown in FIG. 5).

  In some embodiments, the one-arm central-scFv format includes a skew mutation, a pI mutation, and a removal mutation. Accordingly, some embodiments comprise a format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, together with the variable light chain domain of the light chain, the first A first monomer comprising a variable heavy chain domain constituting an Fv that binds to a receptor and an scFv linked to another receptor, b) a skew mutation L368D / K370S, a pI mutation Q295E / N384D / Q418E / N421D, A second monomer comprising a deletion mutation E233P / L234V / L235A / G236del / S267K, and c) a light chain comprising a first variable light chain domain and a constant light chain domain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the one-arm central-scFv format comprises a skew mutation, a pI mutation, a removal mutation, and an FcRn mutation. Accordingly, some embodiments include a format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, and the first of the light chain A first monomer comprising, together with a variable light chain domain, a first variable heavy chain domain comprising an Fv linked to a first receptor and an scFv domain linked to a second receptor, b) a skew mutation A second monomer comprising L368D / K370S, a pI mutation Q295E / N384D / Q418E / N421D, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, and c) a first variable light chain domain And a light chain comprising a constant light chain domain

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

J. et al. 1 arm scFv-mAb
One heterodimeric scaffold specifically used in the present invention is the one-arm scFv-mAb format shown in FIG. 2D. In this embodiment, one monomer contains only the Fc domain and the other monomer uses an scFv domain linked to the N-terminus of the heavy chain, generally by using a linker: vh1-scFv linker- vl1- [arbitrary domain linker] -VH2-CH1-hinge-CH2-CH3 or (in the opposite orientation) vl1-scFv linker-vh1- [arbitrary domain linker] -VH2-CH1-hinge-CH2-CH3. In this format, either Fab moiety is linked to one receptor target and the scFv is linked to another. This embodiment further utilizes a light chain comprising a variable light chain domain and a constant light chain domain, which associate with the heavy chain to form a Fab. For many embodiments herein, these constructs include skew mutations, pI mutations, elimination mutations, additional Fc mutations, etc., as desired and described herein. In this embodiment, suitable Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA, ICOS × TIGIT, TIGIT × ICOS, PD-1 × ICOS, PD-L1 × ICOS, CTLA-4 × ICOS, LAG-3 × ICOS, TIM-3 × ICOS, BTLA × ICOS, OX40 × TIGIT, OX40 × PD-1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BTLA, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG- 3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, ITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA-4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4, 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4-1BB, CTLA-4 × 4-1BB, LAG- 3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB are included (Fab listed first, scFv listed second).

  The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  In addition, Fc domains generally contain skew mutations (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S: S364K, L368E. / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E357L, K370S: S364K / E357Q, T366S / L368A / Y407V: T366W / T366S / L368V / 35 Optionally including a deletion mutation (including that shown in FIG. 6), optionally including a charged scFv linker (including that shown in FIG. 8), and the heavy chain is a pI mutation (as shown in FIG. 5). Including free).

  In some embodiments, the one-arm scFv-mAb format comprises a skew mutation, a pI mutation, and a removal mutation. Thus, some embodiments comprise a format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain of the light chain, A first monomer comprising a first variable heavy chain domain comprising an Fv linked to a first checkpoint inhibitor, and a second variable heavy chain domain, b) skew mutation L368D / K370S, pI mutation N208D / Q295E / N384D / Q418E / N421D, including deletion mutations E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain, linked to a first checkpoint inhibitor as outlined herein A first variable heavy chain domain comprising an Fv, and A second monomer comprising a second variable light chain comprising Fv (ABD) linked to a second checkpoint inhibitor with two variable heavy chains, and c) a first variable light chain domain and a constant A light chain comprising a light chain domain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the one-arm scFv-mAb format comprises a skew mutation, a pI mutation, a removal mutation, and an FcRn mutation. Accordingly, some embodiments include a bottle opener format that includes: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, A first monomer comprising a first variable heavy chain domain and a second variable heavy chain domain comprising an Fv linked to a first checkpoint inhibitor together with one variable light chain domain, b) skew Including the mutation L368D / K370S, the pI mutation N208D / Q295E / N384D / Q418E / N421D, the removal mutation E233P / L234V / L235A / G236del / S267K, the FcRn mutation M428L / N434S, as well as the first variable light chain domain herein. Outlined The first variable heavy chain domain constituting the Fv linked to the first checkpoint inhibitor and the second variable heavy chain constituting the Fv (ABD) linked to the second checkpoint inhibitor together with the second variable heavy chain. A second monomer comprising a variable light chain of: and c) a light chain comprising a first variable light chain domain and a constant light chain domain.

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

K. scFv-mAb format One heterodimeric scaffold specifically used in the present invention is the mAb-scFv format shown in FIG. 2E. In this embodiment, the format relies on the use of scFv N-terminal ligation to one of the monomers, thus forming a third antigen linking domain, wherein the Fab portions of the two monomers are Each links to one target, and “additional” scFv domains link to different targets.

  In this embodiment, the first monomer is in any orientation a first heavy chain (variable heavy) having an N-terminal covalently linked scFv comprising an scFv variable light chain domain, an scFv linker and an scFv variable heavy chain domain. Including (chain domain and constant domain) ((vh1-scFv linker-vl1- [arbitrary domain linker] -vh2-CH1-hinge-CH2-CH3) or (with scFv of opposite orientation) ((vl1-scFv Linker-vh1- [optional domain linker] -vh2-CH1-hinge-CH2-CH3)) The second monomer comprises the heavy chain VH20CH1-hinge-CH2-CH3. Utilizes a common light chain comprising a light chain domain and a constant light chain domain, which associates with the heavy chain and links to one of the target antigens Forms two identical Fabs.For many embodiments herein, these constructs can be skew mutations, pI mutations, removal mutations, additional Fc mutations, as desired and described herein. In this embodiment, suitable Fv pairs include ICOS × PD-1, ICOS × PD-L1, ICOS × CTLA-4, ICOS × LAG-3, ICOS × TIM-3, ICOS × BTLA, ICOS x TIGIT, TIGIT x ICOS, PD-1 x ICOS, PD-L1 x ICOS, CTLA-4 x ICOS, LAG-3 x ICOS, TIM-3 x ICOS, BTLA x ICOS, OX40 x TIGIT, OX40 x PD- 1, OX40 × PD-L1, OX40 × CTLA-4, OX40 × LAG-3, OX40 × TIM-3, OX40 × BT A, TIGIT × OX40, PD-1 × OX40, PD-L1 × OX40, CTLA-4 × OX40, LAG-3 × OX40, TIM-3 × OX40, BTLA × OX40, GITR × TIGIT, GITR × PD-1, GITR × PD-L1, GITR × CTLA-4, GITR × LAG-3, GITR × TIM-3, GITR × BTLA, TIGIT × GITR, PD-1 × GITR, PD-L1 × GITR, CTLA-4 × GITR, LAG-3 × GITR, TIM-3 × GITR, BTLA × GITR, 4-1BB × TIGIT, 4-1BB × PD-1, 4-1BB × PD-L1, 4-1BB × CTLA-4, 4-1BB × LAG-3, 4-1BB × TIM-3, 4-1BB × BTLA, TIGIT × 4-1BB, PD-1 × 4-1BB, PD-L1 × 4 1BB, CTLA-4x4-1BB, LAG-3x4-1BB, TIM-3x4-1BB, and BTLAx4-1BB (Fab listed first, scFv listed second) ).

  The ABD sequences for these combinations may be as disclosed in the sequence listing or figure.

  Furthermore, scFv-mAb format Fc domains generally contain skew mutations (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S: S364K; L368E / K370S: S364K; T411T / E360E / Q362E: D401K; L368D / K370S: S364K / E357L; K370S: S364K / E357Q; / S354C, optionally including a deletion mutation (including that shown in FIG. 6), optionally including a charged scFv linker (including that shown in FIG. 8), Includes pI mutations (including those shown in FIG. 5).

  In some embodiments, the mAb-scFv format includes a skew mutation, a pi mutation, and a removal mutation. Accordingly, some embodiments include a bottle opener format comprising: a) a skew variable S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, and the first variable light chain domain of the light chain And a first monomer comprising a first variable heavy chain domain and a second variable heavy chain domain comprising an Fv linked to a first checkpoint inhibitor, b) skew mutation L368D / K370S, pI A first checkpoint inhibitor as outlined herein, comprising a mutation N208D / Q295E / N384D / Q418E / N421D, a deletion mutation E233P / L234V / L235A / G236del / S267K, together with a first variable light chain domain First variable heavy chain constituting Fv linked to A second monomer comprising a main, and a second variable light chain comprising Fv (ABD) linked to a second checkpoint inhibitor together with a second variable heavy chain, and c) a first variable light A light chain comprising a chain domain and a constant light chain domain. In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In some embodiments, the mAb-scFv format comprises a skew mutation, a pi mutation, a removal mutation, and an FcRn mutation. Accordingly, some embodiments include a bottle opener format that includes: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, an FcRn mutation M428L / N434S, A first monomer comprising a first variable heavy chain domain and a second variable heavy chain domain comprising an Fv linked to a first checkpoint inhibitor together with one variable light chain domain, b) skew Including the mutation L368D / K370S, the pI mutation N208D / Q295E / N384D / Q418E / N421D, the removal mutation E233P / L234V / L235A / G236del / S267K, the FcRn mutation M428L / N434S, as well as the first variable light chain domain herein. Outlined The first variable heavy chain domain constituting the Fv linked to the first checkpoint inhibitor and the second variable heavy chain constituting the Fv (ABD) linked to the second checkpoint inhibitor together with the second variable heavy chain. A second monomer comprising a variable light chain of: and c) a light chain comprising a first variable light chain domain and a constant light chain domain.

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

L. Dual scFv format The present invention also provides a dual scFv format as known in the art and as shown in FIG. 2B. In this embodiment, heterodimeric bispecific antibodies are made from two scFv-Fc monomers (both (vh-scFv linker-vl- [arbitrary domain linker] -CH2-CH3 ) Format or (vl-scFv linker-vh- [arbitrary domain linker] -CH2-CH3) format, or one monomer in one orientation and the other in the other orientation.

  In this case, all ABDs are in scFv format, ICOS and PD-1, ICOS and CTLA-4, ICOS and LAG-3, ICOS and TIM-3, ICOS and PD-L1, ICOS and BTLA, GITR and PD- 1, GITR and CTLA-4, GITR and LAG-3, GITR and TIM-3, GITR and PD-L1, GITR and BTLA, OX40 and PD-1, OX40 and CTLA-4, OX40 and LAG-3, OX40 TIM-3, OX40 and PD-L1, OX40 and BTLA, 4-1BB and PD-1, 4-1BB and CTLA-4, 4-1BB and LAG-3, 4-1BB and TIM-3, 4-1BB PD-L1 and any combination of 4-1BB and BTLA are useful. The ABD sequences for these combinations may be as disclosed in the sequence listing or as shown in the figure.

  Furthermore, dual scFv format Fc domains generally contain skew mutations (eg, a set of amino acid substitutions as shown in FIG. 4), and particularly useful skew mutations are S364K / E357Q: L368D / K370S, L368D / K370S. : S364K, L368E / K370S: S364K, T411T / E360E / Q362E: D401K, L368D / K370S: S364K / E357L, K370S: S364K / E357Q, T366S / L368A / Y367V: T366S / T366 / 36 Selected from the group consisting of S354C, optionally including a deletion mutation (including that shown in FIG. 6), optionally including a charged scFv linker (including that shown in FIG. 8), Includes pI mutations (including those shown in FIG. 5).

  In some embodiments, the dual scFv format includes a skew mutation, a pi mutation, and a removal mutation. Accordingly, some embodiments include a format comprising: a) a skew mutation S364K / E357Q, a removal mutation E233P / L234V / L235A / G236del / S267K, and a first scFv linked to a first receptor. Monomers (VH1-scFv linker-VL1- [optional domain linker-CH2-CH3 or VL1-scFv linker-VH1- [optional domain linker] -CH2-CH3), and b) skew mutation L368D / K370S, A first monomer comprising a deletion mutation E233P / L234V / L235A / G236del / S267K and an scFv linked to a first receptor (VH1-scFv linker-VL1- [optional domain linker] -CH2-CH3 or VL1 -ScFv Anchor -VH1- [any of the domain linker] -CH2-CH3). The pi mutation can be as outlined herein, but the most common is a charged scFv linker with each monomer having an opposite charge. FcRn mutations, particularly 428L / 434S, can optionally be included. In some specific embodiments having these mutations in this format:

  (1) The format includes an ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes an ICOS ABD of H0.66_L0 combined with ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes ABD 1G6_L1.194_H1.279 linked to PD-1, and Fv links to GITR.

  (6) The format includes ABD 1G6_L1.194_H1.279 linked to PD-1, and Fv links to OX40.

  (7) The format includes ABD 1G6_L1.194_H1.279 linked to PD-1, and ABD links to ICOS.

  8) The format includes ABD 1G6_L1.194_H1.279 linked to PD-1, and ABD links to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

M.M. Non-heterodimeric bispecific antibodies As will be appreciated by those skilled in the art, the Fv sequences outlined herein are also monospecific antibodies (eg, “conventional monoclonal antibodies”) or non-heterobispecifics. It can be used in both specificity formats (see FIGS. 2J, K, and L).

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

  Suitable non-heterodimeric bispecific formats are known in the art and are described in Spiss et al. , Spiss et al. , Molecular Immunology (67): 95-106 (2015) and Kontermann, mAbs 4: 2, 182-197 (2012), including several different formats, both of which are by reference In particular, the figures, illustrations and citations for the formats therein are expressly incorporated.

N. DVD-Ig Format In some embodiments, bispecific antibodies generally have a “dual variable domain-Ig” or “DVD-Ig (as described in US Pat. No. 7,612,181). Trademark) ”format (see FIG. 2L), the entire contents of which are hereby incorporated by reference, particularly with respect to the figures and illustrations therein. In the DVD-Ig format, the antibody is tetravalent and bispecific and contains four chains, two homodimeric body weight chains and two identical light chains. Each heavy chain has a VH1- (arbitrary linker) -VH2-CH1-hinge-CH2-CH3 structure, and the two light chains each have a VL1-arbitrary linker-VL2-CL structure, and VH1 and VL1 Forms a first ABD and VH2 and VL2 form a second ABD, where the first and second ABDs are linked to costimulatory and checkpoint receptors. In this embodiment, the appropriate combinations are ICOS and PD-1, ICOS and CTLA-4, ICOS and LAG-3, ICOS and TIM-3, ICOS and PD-L1, ICOS and BTLA, GITR and PD-1, GITR and CTLA-4, GITR and LAG-3, GITR and TIM-3, GITR and PD-L1, GITR and BTLA, OX40 and PD-1, OX40 and CTLA-4, OX40 and LLA-3, OX40 and TIM- 3. OX40 and PD-L1, OX40 and BTLA, 4-1BB and PD-1, 4-1BB and CTLA-4, 4-1BB and LAG-3, 4-1BB and TIM-3, 4-1BB and PD- L1, and 4-1BB and BTLA are included.

  DVD-Ig ™ and central-scFv2 are two formats that are bispecific and tetravalent and thus do not link to costimulatory receptors in a monovalent manner. An exemplary DVD-Ig ™ construct is shown in FIG.

  In some specific embodiments having these mutations in this format:

  (1) The format includes an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes an ICOS ABD of H0.66_L0 combined with ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes ABD 1G6_L1.194_H1.279 linked to PD-1 and ABD linked to GITR.

  (6) The format includes ABD 1G6_L1.194_H1.279 linked to PD-1 and ABD linked to OX40.

  (7) The format includes ABD 1G6_L1.194_H1.279 linked to PD-1 and ABD linked to ICOS.

  8) The format includes ABD 1G6_L1.194_H1.279 linked to PD-1 and ABD linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

O. Trident Format In some embodiments, the bispecific antibodies of the invention are generally in a “trident” format as described in WO2015 / 184203, the entire contents of which, in particular, the figures, illustrations, definitions, and “ The sequences of “heterodimer facilitating domains” or “HPD”, including the “K-coil” and “E-coil” sequences, are incorporated herein by reference. Trident relies on two different HPDs to sulfur as a structural component that associate to form a heterodimeric structure, see FIG. 2M. In this embodiment, the trident format comprises “conventional” heavy and light chains (eg, VH1-CH1-hinge-CH2-CH3 and VL1-CL), a first “diabody-type linking domain” or “DART ( The third chain comprising "registered trademark", VH2- (linker) -VL3-HPD1, and the fourth chain comprising the second DART®, VH3- (linker)-(linker) -VL2-HPD2 including. VH1 and VL1 form a first ABD, VH2 and VL2 form a second ABD, and VH3 and VL3 form a third ABD. As shown in FIG. 2M, in some cases, the second and third ABDs are linked to the same antigen, in this example generally a checkpoint receptor, for example, bivalently, ABD is monovalently linked to costimulatory receptors. In this embodiment, the appropriate combinations are ICOS and PD-1, ICOS and CTLA-4, ICOS and LAG-3, ICOS and TIM-3, ICOS and PD-L1, ICOS and BTLA, GITR and PD-1, GITR and CTLA-4, GITR and LAG-3, GITR and TIM-3, GITR and PD-L1, GITR and BTLA, OX40 and PD-1, OX40 and CTLA-4, OX40 and LLA-3, OX40 and TIM- 3. OX40 and PD-L1, OX40 and BTLA, 4-1BB and PD-1, 4-1BB and CTLA-4, 4-1BB and LAG-3, 4-1BB and TIM-3, 4-1BB and PD- L1, and 4-1BB and BTLA are included.

  In some specific embodiments having these mutations in this format:

  (1) The format includes a Fab ABD linked to an ICOS having an ABD of [ICOS] H0.66_L0.

  (2) The format includes an ICOS ABD of H0.66_L0 combined with an scFv ABD that includes ABD 1G6_L1.194_H1.279 linked to PD-1.

  (3) The format includes ICOS ABD of H0.66_L0 combined with scFv including ABD H3.23_L0.129 linked to CTLA-4.

  (4) The format is ABD 7G8_H. Contains an ICOS ABD of H0.66_L0 combined with 303_L1.34.

  (5) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to GITR.

  (6) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to OX40.

  (7) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1, and Fab links to ICOS.

  8) The format includes scFv including ABD 1G6_L1.194_H1.279 linked to PD-1 and Fab linked to 4-1BB.

  (9) The format includes H0.66_L0 ABD linked to ICOS and ABD linked to PD-L1.

  In this format, specific ABDs linked to human ICOS include [ICOS] _H0L0 and [ICOS] H0.66_L0, and those shown in FIGS. 19, 20, 24, 68, and 77, and Examples include, but are not limited to, those shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665.

  In this format, specific ABDs that link to human GITR include, but are not limited to, those in FIGS. 18, 72, and 73, and those listed in SEQ ID NOs: 26282 to 26290.

  In this format, specific ABDs linked to OX40 include, but are not limited to, those listed in FIGS. 17, 72 and 73, and SEQ ID NOs: 26272-26281.

  In this format, specific ABDs linked to human 4-1BB include, but are not limited to, FIGS. 16, 72, and 73, and SEQ ID NOs: 26262-2671.

  In this format, specific ABDs linked to human PD-L1 include, but are not limited to, FIGS. 15, 73, and 78, and SEQ ID NOs: 3961-4432.

P. Monospecific, Monoclonal Antibodies As will be appreciated by those skilled in the art, the novel Fv sequences outlined herein are also monospecific antibodies (eg, “conventional monoclonal antibodies”) or non-heterobispecific Can be used in both sex formats. Thus, the present invention generally has six CDRs from the figure having IgG1, IgG2, IgG3, or IgG4 constant regions, IgG1, IgG2, and IgG4 (including an IgG4 constant region comprising an amino acid substitution of S228P) and / or Monoclonal (monospecific) antibodies comprising vh and vl sequences are provided and are particularly used in some embodiments. That is, any sequence herein with an “H_L” designation can be linked to the constant region of a human IgG1 antibody.

VIII. Antigen-binding domain (ABD) for target antigen
The bispecific antibodies of the present invention generally comprise two different target receptor antigens (“target pairs”) in either a bivalent or trivalent bispecific format, as shown in FIG. ) Have two different antigen-binding domains (ABD).

  Bispecific antibodies link to a first target antigen that includes a checkpoint receptor and a second target antigen that includes a costimulatory receptor. Suitable checkpoint receptors outlined herein include PD-1, PD-L1, LAG-3, TIM-3, CTLA-4, BTLA, and TIGIT. Suitable costimulatory receptors outlined herein include ICOS, GITR, OX40, and 4-1BB. As outlined

  Suitable target checkpoint antigens include human (and possibly cynomolgus) PD-1, CTLA-4, TIM-3, LAG-3, TIGIT, and BTLA (sequences in the sequence listing). Thus, suitable bispecific antibodies are ICOS and PD-1, ICOS and CTLA-4, ICOS and LAG-3, ICOS and TIM-3, ICOS and PD-L1, ICOS and BTLA, ICOS and TIGIT, GITR And TIGIT, GITR and PD-1, GITR and CTLA-4, GITR and LAG-3, GITR and TIM-3, GITR and PD-L1, GITR and BTLA, OX40 and PD-1, OX40 and TIGIT, OX40 and CTLA -4, IC OX40 OS and LAG-3, OX40 and TIM-3, OX40 and PD-L1, OX40 and BTLA, 4-1BB and PD-1, 4-1BB and CTLA-4, 4-1BB and LAG-3 4-1BB and TIM-3, 4-1BB and PD-L1, TIGIT and 4-1BB, and 4-1BB and B Connect to TLA.

  In general, these bispecific antibodies are referred to as “anti-PD-1 × anti-CTLA-4” or generally “PD-1 × CTLA− for each pair for simplicity or simplicity (and thus interchangeable). 4 "etc. Unless specified herein, the order of enumeration of antigens in the name does not confer structure, ie, PD-1 × CTLA-4 bottle opener antibody may link scFv to PD-1 or CTLA-4. Yes, but in some cases the order identifies the structure as shown.

  As outlined more fully herein, these ABD combinations can be in various formats, as outlined below, generally a combination where one ABD is in the Fab format and the other is in the scFv format. It is. As discussed herein and shown in FIG. 2, some formats use a single Fab and a single scFv (A, C, and D), and some formats have two Fab and A single scFv (E, F, G, H, and I) is used.

A. Antigen-binding domains As discussed herein, the bispecific checkpoint heterodimeric antibodies of the present invention comprise two antigen-binding domains (ABD) each linked to a different checkpoint protein. As outlined herein, these heterodimeric antibodies can be bispecific and bivalent (each antigen linked by a single ABD, for example in the format shown in FIG.
FIG. 1 is a compilation of bladder cancer, breast cancer, colon cancer, head and neck cancer, kidney cancer, lung adenocarcinoma, lung squamous cell carcinoma, ovarian cancer, pancreatic cancer, prostate cancer, and black, edited from The Cancer Genome Atlas (TCGA) Presents PD-1 and T cell costimulatory receptor expression data (RNAseq V2 RSEM) for tumor cancer. The square of the Pearson correlation coefficient was calculated for PD-1 for T cell costimulatory receptors.

  Figures 2A-O show several formats for the bispecific antibodies of the invention. The first is a “bottle opener” format having first and second anti-antigen linking domains. Furthermore, the formats of mAb-Fv, mAb-scFv, center-scFv, center-Fv, center of one arm-scFv, one scFv-mAb, scFv-mAb, dual scFv are all shown. FIG. 2J shows a “center-scFv2” format with two Fab-scFv arms, where the Fab is linked to the first antigen and the scFv is linked to the second antigen. FIG. 2K shows a bispecific mAb format with a first Fab arm linking to the first antigen and a second Fab arm linking to the second antigen. FIG. 2L shows the format of DVD-IgG (see, eg, US Pat. No. 7,612,181, which is expressly incorporated herein by reference and discussed below). FIG. 2M shows a trident format (see, eg, WO2015 / 184203, which is expressly incorporated herein by reference and discussed below). For all the scFv domains shown, they can be either variable heavy chain- (optional linker) -variable light chain, or vice versa, from N-terminus to C-terminus. Furthermore, in 1-arm scFv-mAb, scFv can be linked to either the N-terminus of the heavy chain monomer or the N-terminus of the light chain.

FIG. 3 shows a bottle with ICOS on the Fab side ([ICOS] _H0.66_L0), PD-1 as scFV (1G6_L1.94_H1.279) and containing the M428L / 434S mutation to prolong serum half-life. The arrangement of XENP23104 which is an opener format is shown. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers, including uncharged or negatively charged linkers shown in FIG. 8), and slashes indicate variable domain boundaries. Furthermore, the naming convention indicates the orientation of the scFv from the N-terminus to the C-terminus, and some of the sequences herein are oriented with VH-scFv linker-VL (from N to the C-terminus), while Some are oriented with VL-scFv linker-VH (from N to C-terminus), but as will be appreciated by those skilled in the art, these sequences are also opposite to those depicted herein. Can be used in orientation. The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are included herein.

  4A-E show useful pairs of heterodimerized mutation sets (including skew and pI mutations). In FIGS. 4C and F there are mutations for which there is no corresponding “monomer 2” mutation, which can be used alone for either monomer, or for example the Fab side of a bottle opener For example, a charged scFv linker suitable for a second monomer that utilizes scFv as the second antigen linking domain can be used. A suitable charged linker is shown in FIG.

  FIG. 5 shows a list of equivalent mutation antibody constant regions and their respective substitutions. pI _ (−) indicates a low pI mutation and pI _ (+) indicates a high pI mutation. These can optionally and independently be combined with other heterodimerization mutations of the invention (and other mutations as outlined herein).

  FIG. 6 shows useful excision mutations that remove the FcγR linkage (often referred to as “knockout” or “KO” mutations). In general, deletion mutations are found in both monomers, but in some cases they may be in only one monomer.

  FIG. 7 shows two particularly useful embodiments of the present invention. Although the “non-Fv” component of this embodiment is shown in FIG. 9A, other formats in FIG. 9 can be used as well.

  FIGS. 8A and B show several charged scFv linkers used in increasing or decreasing the pI of heterodimeric antibodies that utilize one or more scFv as components. (+ H) positive linkers are specifically used herein. A single prior art scFv linker with a single charge is described in Whitlow et al. , Protein Engineering 6 (8): 989-995 (1993). It should be noted that this linker was used to reduce aggregation in scFv and increase proteolytic stability.

  FIG. 9 shows the backbone sequences of several useful bottle opener formats based on human IgG1, without the Fv sequences (eg, vh and vl for scFv and Fab sides). Skeletal 1 of the bottle opener is based on human IgG1 (356E / 358M allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del / S267K contains both strand removal mutations. Skeletal 2 of the bottle opener is based on human IgG1 (356E / 358M allotype), with different skew mutations, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and both E233P / L234V / L235A / G236del / S267K Contains strand removal mutations. Bottle Opener Skeleton 3 is based on human IgG1 (356E / 358M allotype), with different skew mutations, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and both E233P / L234V / L235A / G236del / S267K Contains strand removal mutations. The bottle opener scaffold 4 is based on human IgG1 (356E / 358M allotype) and has different skew mutations, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and both E233P / L234V / L235A / G236del / S267K Contains strand removal mutations. Skeletal 5 of the bottle opener is based on human IgG1 (356D / 358L allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del / S267K contains both strand removal mutations. The skeleton 6 of the bottle opener is based on human IgG1 (356E / 358M allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del / S267K both strand removal mutations, as well as N297A both strand mutations. The skeleton 7 of the bottle opener is identical to 6 except that the mutation is N297S. Alternative formats for the bottle opener scaffolds 6 and 7 can exclude both strand removal mutations E233P / L234V / L235A / G236del / S267K. Skeletal 8 is based on human IgG4, S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and E233P / L234V / L235A / G236del / S267K both chains It includes a deletion mutation as well as a mutation in both strands of S228P (EU numbering, which is S241P in Kabat) that eliminates Fab arm exchange, as is known in the art. In an alternative format for the skeleton 8 of the bottle opener, both E233P / L234V / L235A / G236del / S267K strand removal mutations can be excluded, and skeleton 9 is based on human IgG2 and S364K / E357Q: skew of L368D / K370S. Mutation, including a pI mutation on the Fab side of N208D / Q295E / N384D / Q418E / N421D. Skeletal 10 is based on human IgG2 and contains S364K / E357Q: S368D / K370S skew mutation, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and S267K both chain mutations.

  As understood by those skilled in the art and outlined below, these sequences include one monomer containing scFv (optionally including a charged scFv linker) and another monomer containing a Fab sequence, It can be used with any of the vh and vl pairs outlined herein (eg, vh linked to a “Fab heavy chain” and vl linked to a “constant light chain”). That is, any of the Fv sequences outlined herein for anti-CTLA-4, anti-PD-1, anti-LAG-3, anti-TIM-3, anti-TIGIT, and anti-BTLA are scFv (also optionally charged These can be incorporated in any combination into the backbone of FIG. 37, whether or not as having an scFv linker) or as a Fab. The constant light chain shown in FIG. 9A can be used for all of the constructs in the figure, but the kappa constant light chain may also be substituted.

  Note that these bottle opener scaffolds are used in the center-scFv format of FIG. 1F, where an additional second Fab (vh-CH1) having the same antigenic linkage as the first Fab. And vl-constant light chain) are added to the N-terminus of the “bottle opener side” scFv.

  Included in each of these backbones is 90, 95, 98, and 99% identical to the listed sequence (as defined herein) and / or (to those skilled in the art As can be seen, 1, 2, 3, 4, 5, compared to the “parent” in the figure which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4 based on the backbone) A sequence that includes 6, 7, 8, 9, or 10 additional amino acid substitutions, ie, the listed backbone contains additional amino acid modifications in addition to the skew, pI, and excision mutations contained within the backbone of this figure (Generally amino acid substitutions).

  FIG. 10 shows the sequence of a selected number of anti-PD-1 antibodies. It is important to note that these sequences were generated based on human IgG1 with the excision mutation shown in FIG. 6A (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”) . CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein.

FIG. 11 shows a selected number of PD-1 ABDs, with additional anti-PD-1 ABDs listed as SEQ ID NOs: 1-239, 3125-3144, 4697-7594, and 4697-21810. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers, including uncharged or negatively charged linkers shown in FIG. 8), and slashes indicate variable domain boundaries. Furthermore, the naming convention shows the orientation of scFv from N-terminus to C-terminus, some of the sequences in this figure are oriented as VH-scFv linker-VL (from N to C-terminus), while Are oriented with the VL-scFv linker-VH (from N to C-terminus), but as will be appreciated by those skilled in the art, these sequences are also oriented in the opposite direction to their depiction herein. Can be used in The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format.

FIG. 12 shows a selected number of CTLA-4ABD, with additional anti-CTLA-4ABD listed as SEQ ID NOs: 2393-2414 and 3737-3816. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers, including uncharged or negatively charged linkers shown in FIG. 8), and slashes indicate variable domain boundaries. Furthermore, the naming convention shows the orientation of scFv from N-terminus to C-terminus, some of the sequences in this figure are oriented as VH-scFv linker-VL (from N to C-terminus), while Are oriented with the VL-scFv linker-VH (from N to C-terminus), but as will be appreciated by those skilled in the art, these sequences are also oriented in the opposite direction to their depiction herein. Can be used in The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format.

FIG. 13 shows a selected number of LAG-3 ABD, with additional anti-LAG-3 ABDs listed as SEQ ID NOs: 2415-2604 and 3817-3960. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers including uncharged or negatively charged linkers shown in FIG. 8), and the slash indicates the variable domain> boundary (s). Furthermore, the naming convention shows the orientation of scFv from N-terminus to C-terminus, some of the sequences in this figure are oriented as VH-scFv linker-VL (from N to C-terminus), while Are oriented with the VL-scFv linker-VH (from N to C-terminus), but as will be appreciated by those skilled in the art, these sequences are also oriented in the opposite direction to their depiction herein. Can be used in The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format.

  FIG. 14 shows the sequences of selected numbers of anti-TIM-3 antibodies. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein.

  FIG. 15 shows the sequence of a selected number of anti-PD-L1 antibodies. These sequences were generated based on the human IgG1 backbone with the deletion mutations outlined herein (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats as well It is important to note that can be used. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein.

  FIG. 16 shows the sequence of a prototype anti-4-1BB antibody. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein.

  FIG. 17 shows the sequence of a prototype anti-OX40 antibody. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein.

  FIG. 18 shows the sequence of a prototype anti-GITR antibody. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR position may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein.

  FIG. 19 shows the sequence of a prototype anti-ICOS antibody. These sequences were generated based on the human IgG1 backbone with excision mutations (E233P / L234V / L235A / G236del / S267K, “IgG1_PVA_ / S267K”), but other formats can be used as well. It is important to. CDRs are underlined. The exact identification of the CDR positions may vary slightly depending on the numbering used as shown in Table X, as noted herein and for any sequence containing CDRs herein, so Not only the attached CDRs, but also CDRs contained within VH and VL domains using other numbering systems are also included herein.

  FIG. 20 shows the sequence of an exemplary anti-ICOS Fab. CDRs are underlined and slashes (/) indicate variable region boundaries. The exact identification of the CDR positions may vary slightly depending on the numbering used as shown in Table X and is underlined, as applies to any sequence described herein and including the CDRs herein. CDRs included in VH and VL domains using other numbering systems as well as other CDRs are also included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format. It is important to note that these sequences were generated using a 6-histidine (His6 or HHHHHH) C-terminal tag at the C-terminus of the heavy chain being removed.

  FIG. 21 shows the change in melting temperature from the parent Fab (XENP22050) as determined by DSF of the mutant anti-ICOS Fab modified for melting temperature (Tm) and stability.

  FIG. 22 shows the equilibrium dissociation constant (KD), association rate (Ka), and dissociation rate of a mutant anti-ICOS Fab for a mouse Fc fusion of human ICOS captured on an AMC biosensor as determined by Octet. (Kd) is shown.

  FIG. 23 shows the equilibrium dissociation constant (KD), association rate (KA) of mutant anti-ICOS Fab for biotinylated IgG1 Fc fusion of human ICOS captured on SA biosensor as determined by Octet. And the dissociation rate (Kd).

  FIG. 24 shows an exemplary anti-ICOS scFv sequence. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) 4 linker, but as will be appreciated by those skilled in the art, this linker is , Some of which may be replaced by other linkers including uncharged or negatively charged linkers shown in Figure X), with a slash (/) indicating the boundary of the variable region. The nomenclature indicates the orientation of scFv from N-terminus to C-terminus, some of the sequences in this figure are oriented with VH-scFv linker-VL (from N to C-terminus, see FIG. 24) While some are oriented with the VL-scFv linker-VH (N to C-terminus, see FIG. 24B), as will be appreciated by those skilled in the art, these sequences are also described herein. It can be used in an orientation opposite to their depiction. The exact identification of the CDR positions may vary slightly depending on the numbering used as shown in Table X and is underlined, as applies to any sequence described herein and including the CDRs herein. CDRs included in VH and VL domains using other numbering systems as well as other CDRs are also included herein. Furthermore, for all sequences in the figure, these VH and VL sequences can be used in either scFv format or Fab format. It is important to note that these sequences were generated using a polyhistidine (His6 or HHHHHH) C-terminal tag at the C-terminus of the heavy chain being removed.

  FIG. 25 shows the mutant anti-ICOS scFv modified for melting temperature (Tm) and stability, parent scFv as determined by DSF (XENP24352, VH-scFv linker from N-terminus to C-terminus) The change in melting temperature from the orientation of VL is shown.

  FIG. 26 shows the amino acid sequence of a prototype anti-costim × anti-checkpoint antibody in a bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 27 shows the induction of cytokine secretion by a prototype costim / checkpoint bottle opener in a SEB stimulated PBMC assay.

  FIG. 28 shows SEB-stimulated PBMC after treatment with a test sample indicated as induction of IL-2 secretion in naive (non-SEB stimulation).

  FIG. 29 shows that bispecific antibodies increase specificity, enhance anti-tumor activity, and avoid peripheral toxicity because tumor TIL co-expresses immune checkpoint receptors and costimulatory receptors Schematics related to the benefits of costim x checkpoint blocking bispecific antibodies are shown.

  FIG. 30 shows an exemplary anti-ICOS × anti-PD-1 antibody (XENP20896) double positive cells compared to a one-arm anti-PD-1 antibody (XENP20111) and a one-arm anti-ICOS antibody (XENP20266). Is selectively occupied.

  FIG. 31 shows anti-ICOS × anti-PD-1 2 against A) PD-1 and ICOS double positive T cells and B) PD-1 and ICOS double negative T cells after SEB stimulation of human PBMC. The receptor occupancy of the bispecific antibody (XENP20896), 1-arm anti-ICOS antibody (XENP20266), and 1-arm anti-PD-1 antibody (XENP20111) is shown.

  FIG. 32 shows the amino acid sequence of an exemplary anti-ICOS × anti-PD-1 antibody in bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 33 shows the amino acid sequence of an exemplary anti-ICOS × anti-PD-1 antibody in a bottle opener format containing a mutation of M428L / N434S that results in a longer half-life in serum in one or preferably both Fc domains. . Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed.

  FIG. 34 shows the equilibrium dissociation constant (KD) of a mutant anti-ICOS × anti-PD-1 bispecific antibody against a mouse Fc fusion of human ICOS captured on an AMC biosensor as determined by Octet. The association rate (Ka) and dissociation rate (Kd) are shown.

  FIG. 35 shows the equilibrium of mutant anti-ICOS × anti-PD-1 bispecific antibody against Fc fusion of human and ICOS biotinylated IgG1 captured on SA / SAX biosensor as determined by Octet. The dissociation constant (KD), association rate (KA), and dissociation rate (Kd) are shown.

  FIG. 36 shows cell surface ligation of mutant anti-ICOS × anti-PD-1 bispecific antibody to human T cells in SEB stimulated PBMC assay in two separate experiments shown in A) and B).

  FIG. 37 shows mutant anti-ICOS × anti-PD-1 bispecific antibody, 1-arm anti-ICOS antibody and 1-arm anti-ICOS antibody against PD-1 and ICOS double-positive T cells after SEB stimulation of human PBMC. The receptor occupancy of PD-1 antibody (XENP20111) is shown.

  FIG. 38 shows that the mutant anti-ICOS × anti-PD-1 bispecific antibody promotes secretion of A) IL-2 and B) IFNγ secreted from SEB-stimulated PBMC.

  FIG. 39 shows that the mutant anti-ICOS × anti-PD-1 bispecific antibody promotes secretion of A) IL-2 and B) IFNγ secreted from SEB-stimulated PBMC.

  FIG. 40 shows the concentration of IFNγ on A) day 7 and B) day 11 after transplantation of human PBMC and treatment with the indicated test samples.

  FIG. 41 shows the number of CD45 + cells in mice as determined by flow cytometry at A) 11 days and B) 14 days after transplantation of human PBMC and treatment with the indicated test samples.

  FIG. 42 shows the number of A) CD8 + T cells and B) CD4 + T cells in mice as determined by flow cytometry on day 14 after transplantation of human PBMC and treatment with the indicated test samples (** P ≦ 0.01).

  FIG. 43 shows the change in mouse body weight up to day 14 after transplantation of human PBMC and treatment with the indicated test samples (** P ≦ 0.01).

  FIG. 44 shows the concentration of IFNγ at A) 7 days and B) 14 days after transplantation of human PBMC and treatment with the indicated test samples

  FIG. 45 shows the number of CD45 + cells in mice as determined by flow cytometry on day 14 after transplantation of human PBMC and treatment with the indicated test samples.

  FIG. 46 shows the number of A) CD8 + T cells and B) CD4 + T cells in mice as determined by flow cytometry on day 14 after transplantation of human PBMC and treatment with the indicated test samples.

  FIG. 47 shows the change in mouse body weight up to day 12 and day 15 after transplantation of human PBMC and treatment with the indicated test samples.

  FIG. 48 shows the amino acid sequence of an exemplary anti-ICOS × anti-PD-1 antibody and alternative ICOS ABD in bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 49 shows IL-2 cytokine release assay after SEB stimulation of human PBMC and treatment with alternative anti-ICOS × anti-PD-1 bispecific antibody.

  FIG. 50 shows IL-2 cytokine release assay after SEB stimulation of human PBMC and treatment with alternative anti-ICOS × anti-PD-1 bispecific antibody (as fold induction over bivalent anti-RSV mAb). Indicates.

  FIG. 51 shows AKT phosphorylation in SEB-stimulated purified CD3 + T cells after treatment with anti-ICOS × anti-PD-1 bispecific antibody.

  FIG. 52 shows A) IL-17A, B) IL17F, C) IL-22, D) IL-10 with the indicated test samples upon induction with the indicated bivalent anti-RSV as determined by NanoString. , E) IL-9 and F) IFNγ induction of gene expression induction.

  FIG. 53 shows the fold of mean induction of expression of selected immune response genes by the indicated test sample upon treatment with bivalent anti-RSV mAb as determined by NanoString. The shading intensity corresponds to the magnitude of the magnification change.

  FIG. 54 shows the number of CD45 + cells in mice as determined by flow cytometry on day 14 after transplantation of human PBMC and treatment with the indicated test samples.

  FIG. 55 shows the central scFv of FIG. 2F, the central-scFv2 format of FIG. 2J, the bispecific mAb of FIG. 2K, the DVD-Ig of FIG. 2L, and the trident format of FIG. The sequence of some additional format “backbone” positions (eg, without Fv) is shown. In FIG. 2L, the DVD-Ig® linker is shown double underlined with other linkers found in WO2007 / 024715, the entirety of which is specifically incorporated herein by reference, particularly for those sequences. Embedded in the book. In the trident format, other trident linkers and coil-coil arrangements are shown in WO 2015/184203, which is hereby incorporated by reference in its entirety, particularly for those sequences. As will be appreciated by those skilled in the art, bold domains (eg, “VH1”, VH2-scFv linker-VL2 ”, etc.) are separated by a slash“ / ”and may optionally contain any domain linker. All of these scaffolds utilize a kappa constant region in the light chain, although lambda chains can also be used. As outlined herein, with respect to FIGS. 9 and 75, these scaffolds can be combined with any VH and VL domains.

  FIG. 56 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 57 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in a center-scFv format. Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 58 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in a center-scFv2 format. Antibodies are named using the Fab variable region first, scFv variable region second, followed by the chain designation (heavy or light chain). CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 59 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in a bispecific mAb format. The antibody comprises a first Fab variable region for a first antigen, a second Fab variable region for a second antigen, separated by a dash, followed by a chain designation (heavy chain 1 or 1 for the first antigen). Named using light chain 1, heavy chain 2 or light chain 2) for second antigen. CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 60 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in DVD-IgG format. The antibodies are named using a first variable region for a first antigen, a second Fab variable region for a second antigen, followed by a chain designation (heavy or light chain). CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 61 shows the amino acid sequence of an exemplary anti-PD-1 × anti-ICOS antibody in a trident format. Antibodies are separated by dashes using the first VL and VH for the first antigen, including DART, the Fab variable region for the second antigen, followed by the chain designation (heavy or light chain). Named. CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 62 shows the induction of cytokine secretion (IL-2) by alternative format costim × checkpoint blocking bispecific antibody in SEB stimulated PBMC assay.

  FIG. 63 shows the amino acid sequence of an exemplary anti-ICOS × anti-CTLA-4 antibody in bottle opener format (FAB-scFv-Fc). Antibodies are named using the Fab variable region first and scFv variable region second, separated by dashes, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). . CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 64 shows the amino acid sequence of an exemplary anti-LAG-3 × anti-ICOS antibody in a bispecific mAb format. The antibody comprises a first Fab variable region for a first antigen, a second Fab variable region for a second antigen, separated by a dash, followed by a chain designation (heavy chain 1 or 1 for the first antigen). Named using light chain 1, heavy chain 2 or light chain 2) for second antigen. CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 65 shows the amino acid sequence of an exemplary anti-TIM-3 × anti-ICOS antibody in a bispecific mAb format. The antibody comprises a first Fab variable region for a first antigen, a second Fab variable region for a second antigen, separated by a dash, followed by a chain designation (heavy chain 1 or 1 for the first antigen). Named using light chain 1, heavy chain 2 or light chain 2) for second antigen. CDRs are underlined and slashes indicate variable region boundaries. Each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 66 shows the amino acid sequences of anti-ICOS × anti-PD-L1 antibody in bottle opener format (FAB-scFv-Fc) and central-scFv2 format. The bottle openers are named using the Fab variable region first and scFv variable region second, separated by a dash, followed by the chain designation (Fab-Fc heavy chain, scFv-Fc heavy chain or light chain). The Center-scFv2 is named using Fab variable region first, scFv variable region second, followed by chain designation (heavy chain or light chain). CDRs are underlined and slashes indicate variable region boundaries. CDRs are underlined and slashes indicate variable region boundaries. As shown, the scFv domain has either a different orientation of VH-scFv linker-VL or VL-scFv linker-VH (from N-terminus to C-terminus), but this can be reversed. Furthermore, each sequence outlined herein can include or exclude mutations in M428L / N434S that result in a longer half-life in serum in one or preferably both Fc domains.

  FIG. 67 shows the induction of cytokine secretion (IL-2) by additional costim × checkpoint blocking bispecific antibodies in the SEB stimulated PBMC assay.

  FIG. 68 shows the amino acid sequence of an exemplary 1-arm anti-ICOS Fab-Fc antibody. CDRs are underlined and slashes indicate variable region boundaries. Since one amino acid chain is only the Fc domain and the other side is the anti-ICOS Fab side, these are referred to as “one arm” or “one arm” format. The Fc domain contains the S364K / E357Q skew mutation, as well as the pI (−)-equivalent_A mutation shown in FIG. The Fab Fc domain contains the L368D / K370S skew mutation and the pI ISO (+ RR) mutation shown in Figure X. Both Fc domains contain a deletion mutation (E233P / L234V / L235A / G236del / S267K).

  FIG. 69 shows the equilibrium dissociation constant (KD), association rate (Ka) of a mutant 1-arm anti-ICOS Fab-Fc antibody against a mouse Fc fusion of human ICOS captured on an AMC biosensor as determined by Octet. ) And dissociation rate (Kd).

  FIG. 70 shows AKT phosphorylation in SEB-stimulated purified CD3 + T cells after treatment with bivalent and monovalent anti-PD-1 antibody and anti-ICOS × anti-PD-1 bispecific antibody.

  FIG. 71 shows AKT phosphorylation in purified CD3 + T cells after treatment with monovalent anti-ICOS Fab-Fc antibody with alternative anti-ICOS ABD.

  FIG. 72 shows prototype bispecific antibodies (OX40 × PD-1, GITR × PD-1, 4-1BB × PD-1, CTLA-4 × ICOS).

  FIG. 73 shows several prototype mAbs (4-1BB, OX40, GITR, ICOS, PD-L1, and PD-1), whose Fv is combined with other Fvs of the present invention and in any format (Bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one-arm middle-scFv, one-arm scFv-mAb, dual scFv, DVD-Ig, or trident) Can be used. Similarly, some additional ICOS × PD-L1 bottle opener arrays are also shown.

  FIG. 74 shows a further PD-1 × ICOS bottle opener, where in some cases PD-1Fv is in Fab format, ICOS Fv is in scFv format, and in other cases it is reversed.

  FIGS. 75A-D show the sequences of the mAb-scFv backbone for use in the present invention with the addition of the Fv sequences of the present invention. Skeletal 1 of mAb-scFv is based on human IgG1 (356E / 358M allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / It contains a deletion mutation in both strands of L235A / G236del / S267K. Skeletal 2 is based on human IgG1 (356D / 358L allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del Includes removal mutations in both strands of S267K. Skeletal 3 is based on human IgG1 (356E / 358M allotype), S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and E233P / L234V / L235A / G236del Includes both strand removal mutations in S267K, as well as mutations in both strands of N297A. Skeletal 4 is identical to 3 except that the mutation is N297S. Alternative formats for mAb-scFv scaffolds 3 and 4 can exclude both strand removal mutations E233P / L234V / L235A / G236del / S267K. Skeletal 5 is based on human IgG4, S364K / E357Q: Skew mutation of L368D / K370S, N208D / Q295E / N384D / Q418E / N421D Fab pI mutation, and E233P / L234V / L235A / G236del / S267K both chains Including a mutation in both strands of S228P (EU numbering, which is S241P in Kabat) that eliminates Fab arm exchange, as known in the art, and backbone 6 is based on human IgG2 In addition, the S364K / E357Q: L368D / K370S skew mutation and the N208D / Q295E / N384D / Q418E / N421D pI mutation are included on the Fab side. Skeletal 7 is based on human IgG2 and contains S364K / E357Q: S368D / K370S skew mutation, N208D / Q295E / N384D / Q418E / N421D Fab side pI mutation, and S267K both chain mutations.

  As understood by those of skill in the art and outlined below, these sequences are composed of one monomer containing both Fab and scFv (optionally including a charged scFv linker) and another monomer containing the Fab sequence. Can be used with any of the vh and vl pairs outlined herein (eg, vh linked to a “Fab heavy chain” and vl linked to a “constant light chain”). That is, any of those outlined herein for anti-CTLA-4, anti-PD-1, anti-LAG-3, anti-TIM-3, anti-TIGIT, anti-BTLA, anti-ICOS, anti-GITR, country OX40, and anti-4-1BB The Fv sequences can be incorporated in any combination into this FIG. 75 backbone, whether or not as scFv (also with a charged scFv linker optionally) or Fab. Monomer 1 side is the pI negative side of Fab-scFv, heterodimerization mutation L368D / K370S, equivalent pI mutation N208D / Q295E / N384D / Q418E / N421D, removal mutation E233P / L234V / L235A / G236del / S267K (All compared to IgG1). The monomer 2 side is the pI positive side of scFv and contains the heterodimerization mutation 364K / E357Q. However, other skew mutation pairs, in particular, [S364K / E357Q: L368D / K370S], [L368D / K370S: S364K], [L368E / K370S: S364K], [T411T / E360E / Q362E: D401K], [L368D / K370S: S364K / E357L], [K370S: S364K / E357Q], [T366S / L368A / Y407V: T366W], and [T366S / L368A / Y407V / Y394C: T366W / S354C].

  The constant light chain shown in FIG. 75A can be used for all of the constructs in the figure, but the kappa constant light chain may also be substituted.

  These mAb-scFv scaffolds are similar in both the scFv-mAb format of FIG. 1H in the mAb-Fv format (where one monomer contains vl at the C-terminus and the other contains vh at the C-terminus). It should be noted that it is used in one of the monomers with the addition of a scFv domain at the C-terminus.

  Included in each of these backbones is 90, 95, 98, and 99% identical to the listed sequence (as defined herein) and / or (to those skilled in the art As can be seen, 1, 2, 3, 4, 5, compared to the “parent” in the figure which already contains some amino acid modifications compared to the parent human IgG1 (or IgG2 or IgG4 based on the backbone) A sequence that includes 6, 7, 8, 9, or 10 additional amino acid substitutions, ie, the listed backbone contains additional amino acid modifications in addition to the skew, pI, and excision mutations contained within the backbone of this figure (Generally amino acid substitutions).

  FIG. 76 shows several prior art sequences of Fv linked to human PD-1 as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD.

  FIG. 77 shows some prior art sequences of Fv linked to human ICOS as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fv can be associated with checkpoint receptors (eg, PD-1, PD-L1, CTLA-4, TIM-3, LAG-3, TIGIT, and BTLA). Can be combined with linking Fvs (including Fv sequences included herein) and in any format (bottle opener, mAb-Fv, mAb-scFv, center-scFv, bispecific mAb, center-Fv 1 arm center-scFv, 1 arm scFv-mAb, dual scFv, DVD-Ig, or trident). In particular, they can be combined with PD-1 ABD having identifiers 1G6_H1.279_L1.194, 1G6_H1.280_L1.224, 1G6_L1.194_H1.279, 1G6_L1.210_H1.288, and 2E9_H1L1.

  FIG. 78 shows several prior art sequences of Fv linked to human PD-L1 as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD.

  FIG. 79 shows several prior art sequences of Fv linked to human CTLA-4 as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD.

  FIG. 80 shows several prior art sequences of Fv linked to human LAG-3 as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD.

  FIG. 81 shows several prior art sequences of Fv linked to human TIM-3 as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD.

  FIG. 82 shows several prior art sequences of Fv linked to human BTLA as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD.

  FIG. 83 shows some prior art sequences of Fv linked to human TIGIT as VH and VL sequences. As will be appreciated by those skilled in the art, any of these Fvs includes an Fv that is linked to a costimulatory receptor (eg, ICOS, GITR, OX40, or 4-1BB) (the Fv sequences included herein). ) And any format (bottle opener, mAb-Fv, mAb-scFv, middle-scFv, bispecific mAb, middle-Fv, one arm middle-scFv, one arm scFv-mAb, dual scFv, DVD-Ig, or Trident). In particular, they can be combined with ICOS] _H0L0 and [ICOS] H0.66_L0 ABD.

84A-C shows several BTLA ABDs with additional anti-BTLA ABDs listed as SEQ ID NOs: 3705-3736. The CDR is underlined and the scFv linker is double underlined (in the sequence, the scFv linker is a positively charged scFv (GKPGS) ₄ linker, but as will be appreciated by those skilled in the art, this linker is , Some can be replaced by other linkers, including uncharged or negatively charged linkers shown in FIG. 8), and slashes indicate variable domain boundaries. As noted above, the naming convention indicates the orientation of scFv from N-terminus to C-terminus, and in the sequences listed in this figure they are all vh-scFv linker-vl (from N-terminus to C-terminus). Although, these sequences can also be used in the opposite orientation (from N-terminus to C-terminus), vl-linker-vh. The exact identification of CDR positions may vary slightly depending on the numbering used as shown in Table 1, as noted herein and as it applies to any sequence containing CDRs herein, so that the underline is Not only the attached CDRs but also the CDRs contained within the vh and vl domains using other numbering systems are also included herein. Furthermore, for all sequences in the figure, these vh and vl sequences can be used in either scFv format or Fab format.

  FIG. 85 is a matrix of possible combinations of costim and checkpoint ABD, with all possible combinations. “A” in the box means that PD-1 ABD is 1G6_L1.194_H1.279. “B” in the box indicates that ICOS ABD is [ICOS] H. 066_L0. “C” in the box means that PD-1 is the paired scFv. “D” in the box means that CTLA-4 ABD is Fab and is [CTLA-4] _H3_L0.22. “E” in the box means that CTLA-4 ABD is scFv and is [CTLA-4] _H3.23_L0.22. “F” in the box means that LAG-3 ABD is 7G8_H3.30_L1.34. “G” in the box means that BTLA ABD is 9C6_H1.1_L1. “H” in the box means that the combination is a bottle opener. “I” in the box means that the combination is in the center-scFv format.

  FIG. 86 shows two additional ICOS × PD-1 bottle openers.

  A), bispecific trivalent (eg, as depicted in FIG. 2F, G, H, I, or M, one antigen is linked by a single ABD and the other is linked by two ABDs Or bispecific tetravalent (eg, both antigens are linked by two ABDs as depicted in FIGS. 2J and L).

  Further, in general, one of the ABDs includes an scFv as outlined herein in the N-terminal to C-terminal orientation of the vh-scFv linker-vl or vl-scFv linker-vh. According to the format, one or both of the other ABDs generally have a vh domain (generally as a component of a heavy chain) on one protein chain and a vl (generally a component of a light chain) on another protein chain. As a Fab). The “Trident” format uses DART®, which is similar to scFv except that it is in a different orientation and the linker may be slightly longer in general.

  The present invention provides several ABDs that link to several different receptor proteins, as outlined below. As will be appreciated by those skilled in the art, any set of six CDRs or vh and vl domains can be in scFv or Fab format, which is then added to the heavy and light chain constant domains; Here, the heavy chain constant domain includes mutations (including within the CH1 domain and the Fc domain). The scFv sequences included in the sequence listing utilize specific charged linkers, but as outlined herein, use uncharged or other charged linkers, including those shown in FIG. Can do.

  Furthermore, as noted above, the numbering used in the sequence listing for CDR identification is Kabat, but a different numbering can be used, which changes the amino acid sequence of the CDR as shown in Table 1. Will.

  Additional mutations can be made for all of the variable heavy and light chain domains described herein. As outlined herein, in some embodiments, a set of 6 CDRs is 0, 1, 2, 3, 4, or 5 amino acid modifications (amino acid substitutions are specifically used), And a framework (except for CDRs) selected from those listed in FIG. 1 of US Pat. No. 7,657,380, the entire contents of which are incorporated herein by reference. As long as it retains at least about 80, 85, or 90% identity to the cell line sequence, it can have changes in the framework regions of the variable heavy and light chains. Thus, for example, the same CDRs described herein are against a human germline sequence whose framework regions are selected from those listed in FIG. 1 of US Pat. No. 7,657,380. As long as it retains at least 80, 85, or 90% identity, it can be combined with sequences from different frameworks from human germline sequences. Alternatively, a CDR may have amino acid modifications (eg, 1, 2, 3, 4, or 5 amino acid modifications in a set of CDRs (ie, a CDR is a change in a set of 6 CDRs). As long as the total number is less than 6 amino acid modifications, any combination of CDRs can be changed, eg, one change in vlCDR1, two changes in vhCDR2, no change in vhCDR3, etc.)), similarly As long as the region retains at least 80, 85, or 90% identity to a human germline sequence selected from those listed in FIG. 1 of US Pat. No. 7,657,380, the framework Have a change of area.

B. PD-1 Antigen Linking Domain In some embodiments, one of the ABDs links to PD-1. A suitable set of 6 CDRs and / or VH and VL domains, and scFv sequences are shown in FIG. 3, FIG. 11, FIG. -3144, 4697-7594, and 4697-21810, and those in the sequence listing with identifiers 1G6_H1.279_L1.194, 1G6_H1.280_L1.224, 1G6_L1.194_H1.279, 1G6_L1.210_H1.288, and 2E9_H1L1 Contains an array.

  As will be appreciated by those skilled in the art, suitable anti-PD-1 ABDs may be used underlined or using a different numbering scheme as described herein and shown in Table 1. If so, the CDRs identified using other alignments within the vh and vl sequences of these sequences may include these sequences and a set of 6 CDRs as depicted in the figure. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv for PD-1, it is the scFv monomer that is linked to PD-1. As discussed herein, when PD-1 is one of the antigens, the other of the target pair is ICOS (appropriate sequences are FIG. 19, FIG. 20, FIG. 24, FIG. 68, and FIG. 77, and SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665 (these can be scFv sequences, sets of CDR sequences, or vh and vl sequences) ), GITR, OX40, and 4-1BB.

  Particularly useful ABDs that link to human PD-1 include, but are not limited to, 1G6_H1.279_L1.194, 1G6_H1.280_L1.224, 1G6_L1.194_H1.279, 1G6_L1.210_H1.288, and 2E9_H1L1.

  In addition, useful VH and VL sequences linked to human PD-1 are shown in FIG.

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for PD-1, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form an ABD for PD-1, the present invention provides mutant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the variant vh and vl domains are still ABD when measured in at least one of Biacore, Surface Plasmon Resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. The

  Certain preferred embodiments include the Fv of 1G6_L1.194_H1.279 anti-PD-1 in scFv format and are included within any of the bottle opener format skeletons of FIG.

  Certain preferred embodiments include the 1G6_L1.194_H1.279 anti-PD-1 Fv in scFv format and are included within any of the format backbones of FIG.

  Certain preferred embodiments include a 1G6_L1.194_H1.279 anti-PD-1 Fv in scFv format and are included within any of the mAb-scFv format backbones of FIG.

  As shown in FIG. 76 in any format shown in FIG. 2, other embodiments utilize any anti-PD-1 VH and VL domain pair (as either scFv or Fab).

C. CTLA-4 antigen-binding domain In some embodiments, one of the ABDs is linked to CTLA-4. A suitable set of 6 CDR and / or vh and vl domains, and scFv sequences are shown in SEQ ID NOs: 2393-2414 and 3737-3816, and sequences of particular interest in some embodiments are shown in FIGS. 79, and [CTLA-4] _H0.25_L0, [CTLA-4] _H0.26_L0, [CTLA-4] _H0.27_L0, [CTLA-4] _H0.29_L0, [CTLA-4] _H0. 38_L0, [CTLA-4] _H0.39_L0, 0 [CTLA-4] _H0.40_L0, [CTLA-4] _H0.70_L0, [CTLA-4] _H0_L0.22, [CTLA-4] _H2_L0, [CTLA-4 ] _H3.21_L0.124, [CTLA-4] _H3.2 _L0.129, [CTLA-4] _H3.21_L0.132, [CTLA-4] _H3.23_L0.124, [CTLA-4] _H3.23_L0.129, [CTLA-4] _H3.23_L0.132, [CTLA -4] _H3.25_L0.124, [CTLA-4] _H3.25_L0.129, [CTLA-4] _H3.25_L0.132, [CTLA-4] _H3.4_L0.118, [CTLA-4] _H3.4_L0 119, [CTLA-4] _H3.4_L0.12, [CTLA-4] _H3.4_L0.121, [CTLA-4] _H3.4_L0.122, [CTLA-4] _H3.4_L0.123, [CTLA- 4] _H3.4_L0.124, [CTLA-4] _H3.4_L0.125, [CT A-4] _H3.4_L0.126, [CTLA-4] _H3.4_L0.127, [CTLA-4] _H3.4_L0.128, [CTLA-4] _H3.4_L0.129, [CTLA-4] _H3. 4_L0.130, [CTLA-4] _H3.4_L0.131, [CTLA-4] _H3.4_L0.132, [CTLA-4] _H3.5_L2.1, [CTLA-4] _H3.5_L2.2, [CTLA -4] _H3.5_L2.3, [CTLA-4] _H3_L0, [CTLA-4] _H3_L0.22, [CTLA-4] _H3_L0.44, [CTLA-4] _H3_L0.67, and [CTLA-4] _H3_L0 Also included are those sequences in the sequence listing with .74.

As will be appreciated by those skilled in the art, suitable anti-CTLA-4 ABDs may be used underlined or using a different numbering scheme as described herein and shown in Table 1. Where, as a CDR identified using other alignments within the vh and vl sequences outlined herein, these sequences and a set of 6 CDRs as depicted in the figure are May be included. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv for CTLA-4, it is the scFv monomer that is linked to CTLA-4.
As discussed herein, when CTLA-4 is one of the antigens, the other of the target pair is ICOS (appropriate sequences are FIG. 19, FIG. 20, FIG. 24, FIG. 68, and FIG. 77, and SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665, and the identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0), GITR, OX40 , And 4-1BB.

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for CTLA-4, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form an ABD for CTLA-4, the present invention provides mutant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the mutated vh and vl domains are determined by ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can still be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. Is done.

D. TIM-3 antigen binding domain In some embodiments, one of the ABDs links to TIM-3. A suitable set of six CDRs and / or VH and VL domains, and scFv sequences are shown in FIGS. 14 and 81 and SEQ ID NOs: 3345-3704, 4585-4696. ABD sequences that are of particular interest in some embodiments include the identifiers 1D10_H0L0, 1D12_H0L0, 3H3_H1_L2.1, 6C8_H0L0, 6D9_H0_1D12_L0, 7A9_H0L0, 7B11_H0L0, 7B11var_H0L0, and the 7C11 array of L_H0L0 and 0

  As will be appreciated by those skilled in the art, suitable anti-TIM-3 ABDs may be used underlined or using a different numbering scheme as described herein and shown in Table 1. Where the CDRs identified using other alignments within the vh and vl sequences of those shown herein are those CDRs and a set of 6 CDRs as depicted in the figure. May be included. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv to TIM-3, it is the Fab monomer that is linked to TIM-3. As discussed herein, when TIM-3 is one of the antigens, the other of the target pair is the selected ICOS (appropriate sequences are FIGS. 19, 20, 24, 68). , And FIG. 77 and SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665, and the identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0, as shown) GITR, OX40, and 4-1BB.

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for TIM-3, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form the ABD for TIM-3, the present invention provides mutant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the variant vh and vl domains are still ABD when measured in at least one of Biacore, Surface Plasmon Resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. The

LAG-3 antigen binding domain In some embodiments, one of the ABDs links to LAG-3. A suitable set of 6 CDRs and / or vh and vl domains, as well as scFv sequences, are shown in FIGS. 144_L2.142, 2A11_H1_L2.122, 2A11_H1_L2.123, 2A11_H1_L2.124, 2A11_H1_L2.25, 2A11_H1_L2.47, 2A11_H1_L2.50, 2A11_H1_L2.91, 2A11_H1_L11_L11H1, 211 2A11_H4L1, 2A11_H4L2, 7 8_H0L0, 7G8_H1L1, 7G8_H3.18_L1.11, 7G8_H3.23_L1.11, 7G8_H3.28_L1, 7G8_H3.28_L1.11, 7G8_H3.28_L1.13, 7G8_H3.30_1.33_L1.34, 7G8 Including those sequences within.

  As will be appreciated by those skilled in the art, suitable anti-LAG-3 ABDs may be used underlined or using a different numbering scheme as described herein and shown in Table 1. Where a set of 6 CDRs as depicted in these sequences and figures are identified as CDRs identified using other alignments within the vh and vl sequences of the sequences herein. May be included. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv for LAG-3, it is the Fab monomer that is linked to LAG-3. As discussed herein, when LAG-3 is one of the antigens, the other of the target pair is ICOS (appropriate sequences are FIG. 19, FIG. 20, FIG. 24, FIG. 68, and FIG. 77, and SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665, and the identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0), GITR, OX40 , And 4-1BB.

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for LAG-3, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form an ABD for LAG-3, the present invention provides mutant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the mutated vh and vl domains are determined by ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can still be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. Is done.

E. BTLA antigen binding domain In some embodiments, one of the ABDs links to BTLA. A suitable set of 6 CDRs and / or VH and VL domains, and scFv sequences are shown in FIGS. 82, 84, and SEQ ID NOs: 3705-3736, and have the identifiers 9C6_H0L0, 9C6_H1.1_L1, and 9C6_H1.11_L1 These sequences in the sequence listing are also included.

  As will be appreciated by those skilled in the art, suitable anti-BTLA ABDs are either underlined or use different numbering schemes as described herein and shown in Table 1. In some cases, the CDRs identified using other alignments within the vh and vl sequences shown herein may include these sequences and a set of 6 CDRs as depicted in the figure. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv to BTLA, it is the Fab monomer that is linked to BTLA. As discussed herein, when BTLA is one of the antigens, the other of the target pair is ICOS (appropriate sequences are shown in FIGS. 19, 20, 24, 68, and 77, And SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665, and identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0), GITR, OX40, and 4-1BB is selected.

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for BTLA, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form an ABD for BTLA, the present invention provides variant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the mutated vh and vl domains are determined by ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can still be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. Is done.

F. TIGIT antigen-binding domain In some embodiments, one of the ABDs is linked to TIGIT. The appropriate set of 6 CDR and / or VH and VL domains, as well as the scFv sequence are shown in FIG. And SEQ ID NOs: 4433-4585.

  As will be appreciated by those skilled in the art, suitable anti-TIGIT ABDs are either underlined or use a different numbering scheme as described herein and shown in Table 1. In some cases, the CDRs identified using other alignments within the vh and vl sequences shown herein may include these sequences and a set of 6 CDRs as depicted in the figure. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv for TIGIT, it is the Fab monomer that is linked to TIGIT. As discussed herein, when TIGIT is one of the antigens, the other of the target pair is the selected ICOS (appropriate sequences are FIGS. 19, 20, 24, 68, and 77 and shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665, and identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0), GITR, OX40 and 4-1BB.

G. PD-L1 antigen binding domain In some embodiments, one of the ABDs links to PD-L1. A suitable set of 6 CDR and / or VH and VL domains, as well as scFv sequences, are shown in FIGS.

  As will be appreciated by those skilled in the art, suitable anti-TIGIT ABDs are either underlined or use a different numbering scheme as described herein and shown in Table 1. In some cases, the CDRs identified using other alignments within the vh and vl sequences shown herein may include these sequences and a set of 6 CDRs as depicted in the figure. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv for TIGIT, it is the Fab monomer that is linked to TIGIT. As discussed herein, when TIGIT is one of the antigens, the other of the target pair is the selected ICOS (appropriate sequences are FIGS. 19, 20, 24, 68, and 77 and shown in SEQ ID NOs: 27869-28086, 28087-28269, 27193-27335, 28549-28556, and 28557-28665, and identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0), GITR, OX40 and 4-1BB.

H. ICOS antigen binding domain In some embodiments, one of the ABDs links to ICOS. Appropriate set of 6 CDR and / or vh and vl domains, and scFv sequence ICOS (appropriate sequences are FIG. 19, FIG. 20, FIG. 24, FIG. 68 and FIG. -28269, 27193-27335, 28549-28556, and 28557-28665, and those with identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0.

  As will be appreciated by those skilled in the art, suitable anti-ICOS ABDs are either underlined or use different numbering schemes as described herein and shown in Table 1. In some cases, CDRs identified using other alignments within the sequences disclosed herein may include these sequences and a set of 6 CDRs as depicted in the figures. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv to ICOS, it is the Fab monomer that is linked to the ICOS. As discussed herein, when ICOS is one of the antigens, the other of the target pair is PD-1, FIG. 3, FIG. 10, FIG. 11, FIG. 72, FIG. 76, and SEQ ID NOS: 1-239, 3125-3144, 4697-7594, and 4697-21810, and identifiers 1G6_H1.279_L1.194, 1G6_H1.280_L1.224, 1G6_L1.194_H1.279, 1G6_L1.210_H1.288 , And 2E9_H1L1 (which includes those sequences in the sequence listing with scFv sequences, CDR sequence sets or vh and vl sequences), CTLA-4 (suitable sequences are FIGS. 12, 72, and 79, And SEQ ID NOs: 2393-2414 and 3737-3816 (this scFv sequence, may be a CDR sequence set or vh and vl sequences)), TIM-3 (suitable sequences are shown in FIG. 14, FIG. 81 and SEQ ID NOs: 3345-3704, 4585-4696 (this is the scFv sequence, CDR sequences set or vh and vl sequences)), LAG-3 (suitable sequences are shown in FIG. 13, FIG. 80 and SEQ ID NOs: 2415-2604, 3817-3960 (which are scFv sequences, CDR sequences) Can be a set or vh and vl sequences)), BTLA (appropriate sequences are shown in FIGS. 82 and 84 and SEQ ID NO: 3705-3736 (which are scFv sequences, CDR sequence sets, or vh and vl sequences) Possible)), and TIGIT (suitable sequences are shown in Figure XX and SEQ ID NO: 4433- Shown in 585 (which, scFv sequence, possible with CDR sequences set or vh and vl sequences) are selected from).

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for ICOS, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form the ABD for ICOS, the present invention provides mutant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the mutated vh and vl domains are determined by ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can still be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. Is done.

  Certain preferred embodiments include Fab format [ICOS] _H0.66_L0 anti-ICOS Fv and are included within any of the bottle opener format skeletons of FIG.

  Certain preferred embodiments include [ICOS] _H0_L0 anti-ICOS Fv in scFv format and are contained within any of the bottle opener format skeletons of FIG.

  Certain preferred embodiments include [ICOS] _H0.66_L0 anti-ICOS Fv in scFv format and are contained within any of the mAb-scFv format skeletons of FIG.

  Certain preferred embodiments include the Fab format [ICOS] _H0.66_L0 anti-ICOS Fv and are included within any of the mAb-scFv format skeletons of FIG.

  Certain preferred embodiments include [ICOS] _H0.66_L0 anti-ICOS Fv in the Fab format and are included within any of the format skeletons of FIG.

I. GITR antigen binding domain In some embodiments, one of the ABDs links to GITR. A suitable set of 6 CDR and / or vh and vl domains, as well as the scFv sequence FITR (suitable sequences are shown in FIGS. 18, 72 and 73 and SEQ ID NOs: 26282-26290.

  As will be appreciated by those skilled in the art, suitable anti-GITR ABDs are either underlined or use different numbering schemes as described herein and shown in Table 1. In some cases, CDRs identified using other alignments within the sequences disclosed herein may include these sequences and a set of 6 CDRs as depicted in the figures. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv to GITR, it is the Fab monomer that is linked to the GITR. As discussed herein, when GITR is one of the antigens, the other of the target pair is PD-1, FIG. 3, FIG. 10, FIG. 11, FIG. 72, FIG. 73, FIG. 76, and SEQ ID NOS: 1-239, 3125-3144, 4697-7594, and 4697-21810, and identifiers 1G6_H1.279_L1.194, 1G6_H1.280_L1.224, 1G6_L1.194_H1.279, 1G6_L1.210_H1.288 , And 2E9_H1L1 (which includes those sequences in the sequence listing with scFv sequences, CDR sequence sets or vh and vl sequences), CTLA-4 (suitable sequences are FIGS. 12, 72, and 79, And SEQ ID NOs: 2393-2414 and 3737-3816 (this scFv sequence, may be a CDR sequence set or vh and vl sequences)), TIM-3 (suitable sequences are shown in FIG. 14, FIG. 81 and SEQ ID NOs: 3345-3704, 4585-4696 (this is the scFv sequence, CDR sequences set or vh and vl sequences)), LAG-3 (suitable sequences are shown in FIG. 13, FIG. 80 and SEQ ID NOs: 2415-2604, 3817-3960 (which are scFv sequences, CDR sequences) Can be a set or vh and vl sequences)), BTLA (appropriate sequences are shown in FIGS. 82 and 84 and SEQ ID NO: 3705-3736 (which are scFv sequences, CDR sequence sets, or vh and vl sequences) Possible)), and TIGIT (suitable sequences are shown in Figure XX and SEQ ID NO: 4433- Shown in 585 (which, scFv sequence, possible with CDR sequences set or vh and vl sequences) are selected from).

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for GITR, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form the ABD for GITR, the present invention provides mutant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the mutated vh and vl domains are determined by ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can still be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. Is done.

J. et al. OX40 Antigen Linking Domain In some embodiments, one of the ABDs is linked to OX40. A suitable set of 6 CDR and / or vh and vl domains and scFv sequences for OX40 are provided (suitable sequences are shown in FIGS. 17, 72 and 73 and SEQ ID NOs: 26272-26281). It is.

  As will be appreciated by those skilled in the art, suitable anti-OX40 ABDs are either underlined or use different numbering schemes as described herein and shown in Table 1. In some cases, CDRs identified using other alignments within the sequences disclosed herein may include these sequences and a set of 6 CDRs as depicted in the figures. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv for OX40, it is the Fab monomer that is linked to GITR. As discussed herein, when OX40 is one of the antigens, the other of the target pair is PD-1, FIG. 3, FIG. 10, FIG. 11, FIG. 72, FIG. 73, FIG. 76, and SEQ ID NOS: 1-239, 3125-3144, 4697-7594, and 4697-21810, and identifiers 1G6_H1.279_L1.194, 1G6_H1.280_L1.224, 1G6_L1.194_H1.279, 1G6_L1.210_H1.288 , And 2E9_H1L1 (including those sequences in the sequence listing with scFv sequences, CDR sequence sets or vh and vl sequences), CTLA-4 (suitable sequences are FIGS. 12, 72, and 79, And SEQ ID NOs: 2393-2414 and 3737-3816 (this scFv sequence, may be a CDR sequence set or vh and vl sequences)), TIM-3 (suitable sequences are shown in FIG. 14, FIG. 81 and SEQ ID NOs: 3345-3704, 4585-4696 (this is the scFv sequence, CDR sequences set or vh and vl sequences)), LAG-3 (suitable sequences are shown in FIG. 13, FIG. 80 and SEQ ID NOs: 2415-2604, 3817-3960 (which are scFv sequences, CDR sequences) Can be a set or vh and vl sequences)), BTLA (appropriate sequences are shown in FIGS. 82 and 84 and SEQ ID NO: 3705-3736 (which are scFv sequences, CDR sequence sets, or vh and vl sequences) Possible)), and TIGIT (suitable sequences are shown in Figure XX and SEQ ID NO: 4433- Shown in 585 (which, scFv sequence, possible with CDR sequences set or vh and vl sequences) are selected from).

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for OX40, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form an ABD for OX40, the present invention provides mutant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the mutated vh and vl domains are determined by ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can still be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. Is done.

K. 4-1BB Antigen Linking Domain In some embodiments, one of the ABDs links to 4-1BB. Appropriate set of 6 CDR and / or vh and vl domains, and scFv sequence 4-1BB (appropriate sequences are shown in FIG. 16, FIG. 72, and FIG. 73 and SEQ ID NOs: 26262-2671.

  As will be appreciated by those skilled in the art, suitable anti-4-1BB ABDs may be used underlined or using a different numbering scheme as described herein and shown in Table 1. If so, the CDRs identified using other alignments within the sequences disclosed herein may include these sequences and a set of 6 CDRs as depicted in the figures. Suitable ABDs can also include these sequences used as scFv or Fab and the entire sequence of vh and vl as shown in the figure. In many of the embodiments herein that include an Fv for 4-1BB, it is the Fab monomer that is linked to 4-1BB. As discussed herein, when ICOS is one of the antigens, the other of the target pair is PD-1, FIG. 3, FIG. 10, FIG. 11, FIG. 72, FIG. 76, and SEQ ID NOS: 1-239, 3125-3144, 4697-7594, and 4697-21810, and identifiers 1G6_H1.279_L1.194, 1G6_H1.280_L1.224, 1G6_L1.194_H1.279, 1G6_L1.210_H1.288 , And 2E9_H1L1 (which includes those sequences in the sequence listing with scFv sequences, CDR sequence sets or vh and vl sequences), CTLA-4 (suitable sequences are FIGS. 12, 72, and 79, And SEQ ID NOs: 2393-2414 and 3737-3816 (this scFv sequence, may be a CDR sequence set or vh and vl sequences)), TIM-3 (suitable sequences are shown in FIG. 14, FIG. 81 and SEQ ID NOs: 3345-3704, 4585-4696 (this is the scFv sequence, CDR sequences set or vh and vl sequences)), LAG-3 (suitable sequences are shown in FIG. 13, FIG. 80 and SEQ ID NOs: 2415-2604, 3817-3960 (which are scFv sequences, CDR sequences) Can be a set or vh and vl sequences)), BTLA (appropriate sequences are shown in FIGS. 82 and 84 and SEQ ID NO: 3705-3736 (which are scFv sequences, CDR sequence sets, or vh and vl sequences) Possible)), and TIGIT (suitable sequences are shown in Figure XX and SEQ ID NO: 4433- Shown in 585 (which, scFv sequence, possible with CDR sequences set or vh and vl sequences) are selected from).

  In addition to the parent CDR set disclosed in the sequence listing that forms the ABD for 4-1BB, the present invention provides a variant CDR set. In one embodiment, the set of six CDRs is such that the ABD is still the target antigen when measured by at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg, Octet assay) assays. Can have 1, 2, 3, 4 or 5 amino acid changes from the parent CDR, as long as they can be linked to, the latter is particularly used in many embodiments.

  In addition to the parental variable heavy and variable light chain domains disclosed herein that form an ABD for 4-1BB, the present invention provides mutant vh and vl domains. In one embodiment, the mutated vh and vl domains are still ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. Each can have an amino acid change from 1, 2, 3, 4, 5, 6, 7, 8, 9, or 10 from the parental vh and vl domains as long as it can be linked to the target antigen, The latter is particularly used in many embodiments. In another embodiment, the mutated vh and vl domains are determined by ABD when measured in at least one of Biacore, surface plasmon resonance (SPR) and / or BLI (Biolayer Interferometry, eg Octet Assay) assays. As long as it can still be linked to the target antigen, it can be at least 90, 95, 97, 98, or 99% identical to the respective parental vh and vl domains, the latter being particularly used in many embodiments. Is done.

L. Specific Bispecific Embodiments The present invention provides several specific bispecific antibodies as outlined below.

1. ICOS × PD-1
The present invention provides bispecific heterodimeric antibodies that link ICOS and PD-1 each monovalently and optionally both bivalently as outlined herein.

  In one embodiment, the PD-1 ABD is 1G6_L1.194_H1.279 and the ICOS ABD is FIG. 19, FIG. 20, FIG. 24, FIG. 68, and FIG. ~ 27335, 28549-28556, and 28557-28665, and those with identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0.

  In one embodiment, the ICOS × PD-1 bispecific antibody has XENP numbers 20261, 20730, 20896, 22432-22238, 22731-222748, 22878-22894, 22931-22963, 22950-22961, 23090-23093, 23295. -23296, 23301, 23405, 23408, 23410 (all do not have 428L / 434S, but they can have those substitutions), XENP numbers 22730, 22917-22928, 22935-22937, 22974-22979, 22995 ~ 22996, 23001, 23103, 23104 (all have 428L / 434S, but they may not have those substitutions), XENP numbers 23411,218 8, 21829, 21830, 21831, 23348, 23059 (using prior art ICOS arrays), XENP numbers 18920, 24125, 24130 (additional bottle openers), XENP 23406, 23407, 24128 (ICOS × PD-1 center-scFv) XENP24123 (ICOS × PD-1 central scFv2), XENP24134 (ICOS × PD-1 bispecific mAb), 24122 (ICOS × PD-1 DVD-Ig), XENP24132, 24133 (ICOS × PD-1 trident) Selected.

2. ICOS × PD-L1
In this embodiment, the ICOS ABD is shown in FIG. 19, FIG. 20, FIG. 24, FIG. 68, and FIG. Shown with identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0.

3. ICOS x CTLA-4
In this embodiment, the ICOS ABD is shown in FIG. 19, FIG. 20, FIG. 24, FIG. 68, and FIG. Shown with identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0.

  In one embodiment, the bottle opener format with the Fab ICOS ABD of [ICOS] H 0.66_L0 is the scFv CTLA-4 ABD of [CTLA-4] _H3.32_L0.129, particularly in the bottle opener skeleton 1 of FIG. It becomes a pair.

4). ICOS × LAG-3
In this embodiment, the ICOS ABD is shown in FIG. 19, FIG. 20, FIG. 24, FIG. 68, and FIG. Shown with identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0.

5. ICOS × TIM-3
This embodiment is provided by ICOS ABD as shown in FIG. 19, FIG. Shown with identifiers [ICOS] _H0.66_L0 and [ICOS] _H0L0.

M.M. Alloantibodies The present invention further provides antibodies that share amino acid sequence identity with the antibodies outlined herein.

  In one embodiment, bispecific antibodies are made having amino acid variations in one or more CDRs of the Vh and Vl sequences outlined herein and in the figures. In one embodiment, antibodies having 1, 2, 3, 4, or 5 amino acid differences in one or more CDRs of the vh and vl chains outlined herein are provided. These amino acid variations may be present in one CDR or may spread between more than one CDR. These amino acid variations can also be present in one or both of the Fvs of the bispecific antibody, for example, on the ICOS Fv side (generally a Fab but can be a scFv as outlined herein). There may be two amino acid mutations in the CDR and one on the PD-1 side.

  Similarly, the present invention has at least 95, 96, 97, 98, or 99% amino acid identity to the sequences outlined herein, particularly to the variable heavy and / or variable light chain domains. An antibody is provided. This sequence identity may also be present on one Fv or both Fvs of a bispecific antibody. Thus, bispecific antibodies are provided that are 95-99% identical to the variable heavy and / or variable light chain domains outlined in the drawings.

IX. Nucleic acids of the invention The invention further provides nucleic acid compositions encoding the bispecific antibodies of the invention (or nucleic acids encoding them as well in the case of “monospecific” antibodies).

  As will be appreciated by those skilled in the art, the nucleic acid composition will depend on the heterodimeric protein format and scaffold. Thus, for example, when a format requires three amino acid sequences, three nucleic acid sequences can be incorporated into one or more expression vectors for expression. Similarly, some formats (eg, a dual scFv format as disclosed in FIG. 2) require only two nucleic acids, and can similarly be incorporated into one or two expression vectors. Some formats require 4 amino acids (bispecific mAb).

  As is known in the art, a nucleic acid encoding a component of the invention is a host cell used as is known in the art and for producing the heterodimeric antibodies of the invention. Depending on the expression vector. In general, a nucleic acid is operably linked to any number of regulatory elements (promoter, origin of replication, selectable marker, ribosome linking site, inducer, etc.). The expression vector can be an extrachromosomal vector or an integration vector.

  The nucleic acid and / or expression vector of the invention is then transformed into any number of different types of host cells known in the art, including mammalian cells, bacterial cells, yeast cells, insect cells and / or fungal cells. However, mammalian cells (eg, CHO cells) are used in many embodiments.

  In some embodiments, the nucleic acid encoding each monomer and the optional nucleic acid encoding the light chain are each a single expression vector, generally under different or the same promoter control, as applicable depending on the format. Contained within. In embodiments specifically used in the present invention, each of these two or three nucleic acids is contained in a different expression vector. Different vector ratios can be used to drive heterodimer formation, as shown herein and in 62 / 025,931, incorporated herein by reference. That is, surprisingly, the protein is a first monomer: second monomer: light chain (in many embodiments herein having three polypeptides, including heterodimeric antibodies). In a ratio of 1: 1: 2, but these are not the ratios that give the best results.

  The heterodimeric antibody of the present invention is produced by culturing a host cell containing an expression vector, as is well known in the art. Once manufactured, conventional antibody purification steps including ion exchange chromatography steps are performed. As discussed herein, the pi of the two monomers differing by at least 0.5 allows for separation by ion exchange chromatography or isoelectric focusing, or other methods sensitive to isoelectric point Can be. That is, the pI that changes the isoelectric point (pI) of each monomer so that the isoelectric point (pI) of each monomer has a different pI and the heterodimer also has a different pI. Inclusion of substitution facilitates isoelectric focusing (eg, anion exchange column, cation exchange column) of “triple F” heterodimers. These substitutions also help identify and monitor any contaminating dual scFv-Fc and mAb homodimers after purification (eg, IEF gel, cIEF, and analytical IEX columns).

X. Biological and Biochemical Functionality of Heterodimeric Immunomodulatory Antibodies In general, the bispecific immunomodulatory antibodies of the invention are administered to patients with cancer and efficacy is described herein. It is evaluated in several ways. For this reason, standard assays for efficacy, such as assessment of cancer burden, tumor size, presence or extent of metastases, can be performed, but immunooncological treatment is also assessed based on immune status assessment as well. obtain. This can be performed in several ways, including both in vitro and in vivo assays. For example, an assessment of changes in immune status (eg, presence of ICOS + CD4 + T cells after ipi treatment) can be performed in conjunction with “classical” measurements such as tumor burden, size, invasiveness, LN involvement, metastasis, etc. . Thus, any or all of the following can be assessed: CD4 ⁺ T cell activation or proliferation, CD8 ⁺ T (CTL) cell activation or proliferation, CD8 ⁺ T cell mediated cytotoxic activity and / or CTL-mediated cell depletion, the inhibitory effect of PVRIG on NK cell activity and NK-mediated cell depletion, Treg cell differentiation and proliferation and the potentiating effect of PVRIG on Treg cells or bone marrow-derived suppressor cell (MDSC) -mediated immunosuppression or immune tolerance, And / or the effect of PVRIG on inflammatory cytokine production by immune cells, eg, IL-2, IFN-γ or TNF-α production by T cells or other immune cells.

  In some embodiments, treatment assessment is performed by assessing immune cell proliferation using, for example, CFSE dilution, Ki67 intracellular staining of immune effector cells, and 3H-thymidine incorporation.

  In some embodiments, the treatment assessment comprises one or more of cell degranulation as measured by surface expression of CD25, CD69, CD137, ICOS, PD1, GITR, OX40, and CD107A. This is done by assessing increased protein levels of increased or activated related markers.

  In general, gene expression assays are performed as is known in the art.

  In general, protein expression measurements are similarly performed as is known in the art.

In some embodiments, the assessment of treatment may include inferring a number of cellular parameters such as enzyme activity (including protease activity), cell membrane permeability, cell adhesion, ATP production, coenzyme production, and nucleotide uptake activity. This is done by assessing cytotoxic activity as measured by target cell viability detection. Specific examples of these assays include trypan blue or PI staining, ⁵¹ Cr or ³⁵ S release method, LDH activity, MTT and / or WST assay, calcein-AM assay, luminescent assay, and others However, it is not limited to these.

  In some embodiments, assessment of treatment is performed by assessing T cell activity as measured by cytokine production and using well known techniques, IFNγ, TNFα, GM-CSF, IL2, IL6, IL4. , IL5, IL10, IL13 and other cytokines are used to measure either intracellularly or in the culture supernatant.

  Thus, treatment evaluation can be performed using an assay that evaluates one or more of the following: (i) increasing the immune response, (ii) αβ and / or γδ T cells. Increase activation, (iii) increase cytotoxic T cell activity, (iv) increase NK and / or NKT cell activity, (v) reduce αβ and / or γδ T cell suppression (Vi) increasing proinflammatory cytokine secretion; (vii) increasing IL-2 secretion; (viii) increasing interferon-γ production; (ix) increasing Th1 response; (X) reducing the Th2 response, (xi) regulatory T cells (reducing or eliminating at least one cell number and / or activity of Treg) A.

Assays for Measuring Efficacy In some embodiments, T cell activation is assessed using a mixed lymphocyte reaction (MLR) assay, as is known in the art. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, signal transduction pathway assays measure an increase or decrease in immune response as measured, for example, by phosphorylation or dephosphorylation of different factors or by measuring other post-translational modifications. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is measured by, for example, cytokine secretion or by proliferation or by changes in expression of activation markers such as, for example, CD137, CD107a, PD1, etc. The increase or decrease of T cell activation is measured. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is activated, for example, by direct killing of target cells, eg, cancer cells, or by cytokine secretion, or by proliferation, or, eg, CD137, CD107a, PD1, etc. An increase or decrease in cytotoxic T cell activity as measured by a change in marker expression is measured. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, signal transduction pathway assays are measured, for example, by direct killing of target cells, eg, cancer cells, or by cytokine secretion, or by changes in the expression of activation markers such as, eg, CD107a. Measuring an increase or decrease in cellular activity of NK and / or NKT. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is measured by, for example, cytokine secretion, or by proliferation, or by changes in the expression of activation markers such as, for example, CD137, CD107a, PD1, etc. The increase or decrease in T cell suppression is measured. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, signal transduction pathway assays are performed for pro-inflammatory cytokine secretion as measured, for example, by ELISA, by Luminex, or by Multiplex bead-based methods, or by intracellular staining and FACS analysis, or by Alispot et al. Measure the increase or decrease. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is an increase in IL-2 secretion as measured, for example, by ELISA or by Luminex, or by a Multiplex bead-based method, or by intracellular staining and FACS analysis, or by Alispot et al. Or measure the decrease. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is an increase in interferon gamma production as measured, for example, by ELISA or by Luminex, or by Multiplex bead-based methods, or by intracellular staining and FACS analysis, or by Alispot et al. Measure the decrease. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in Th1 response as measured, for example, by cytokine secretion or by changes in activation marker expression. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in Th2 response as measured, for example, by cytokine secretion or by changes in activation marker expression. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in the number and / or activity of at least one regulatory T cell (Treg) as measured, for example, by flow cytometry or by IHC. A decrease in response indicates immunostimulatory activity. The appropriate decrease is the same as for the increase outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in the number of M2 macrophage cells as measured, for example, by flow cytometry or by IHC. A decrease in response indicates immunostimulatory activity. The appropriate decrease is the same as for the increase outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in M2 macrophage tumorigenesis promoting activity as measured, for example, by cytokine secretion or by altered expression of activation markers. A decrease in response indicates immunostimulatory activity. The appropriate decrease is the same as for the increase outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in N2 neutrophil increase as measured, for example, by flow cytometry or by IHC. A decrease in response indicates immunostimulatory activity. The appropriate decrease is the same as for the increase outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in N2 neutrophil tumorigenic activity as measured, for example, by cytokine secretion or by changes in expression of activation markers. A decrease in response indicates immunostimulatory activity. The appropriate decrease is the same as for the increase outlined below.

  In one embodiment, the signal transduction pathway assay is a measure of T cell activation, as measured, for example, by cytokine secretion or by proliferation or by changes in the expression of activation markers such as, for example, CD137, CD107a, PD1, etc. Measure the increase or decrease in inhibition. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is activated, for example, by direct killing of target cells, eg, cancer cells, or by cytokine secretion, or by proliferation or, eg, CD137, CD107a, PD1, etc. The increase or decrease in inhibition of CTL activation is measured as measured by changes in marker expression. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in αβ and / or γδ T cell depletion as measured, for example, by changes in expression of activation markers. A decrease in response indicates immunostimulatory activity. The appropriate decrease is the same as for the increase outlined below.

  In one embodiment, the signal transduction pathway assay is measured by, for example, cytokine secretion, or by proliferation, or by changes in the expression of activation markers such as, for example, CD137, CD107a, PD1, etc. Measure the increase or decrease in T cell response. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay stimulates an antigen-specific memory response, as measured, for example, by cytokine secretion or by proliferation or by changes in the expression of activation markers such as, for example, CD45RA, CCR7, etc. Measure the increase or decrease in. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is measured, for example, by a cytotoxicity assay such as MTT, Cr release, calcine AM, or by a flow cytometry based assay such as CFSE dilution or propidium iodide staining, for example. Measure the increase or decrease in apoptosis or lysis of cancer cells. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is measured, for example, by a cytotoxicity assay such as MTT, Cr release, calcine AM, or by a flow cytometry based assay such as CFSE dilution or propidium iodide staining, for example. The increase or decrease in stimulation of the cytotoxic or cytostatic effect of a cancer cell is measured. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is measured, for example, by a cytotoxicity assay such as MTT, Cr release, calcine AM, or by a flow cytometry based assay such as CFSE dilution or propidium iodide staining, for example. Measure the increase or decrease in direct killing of cancer cells. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay measures an increase or decrease in Th17 activity as measured, for example, by cytokine secretion or by proliferation or by changes in expression of activation markers. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, the signal transduction pathway assay is measured, for example, by a cytotoxicity assay such as MTT, Cr release, calcine AM, or by a flow cytometry based assay such as CFSE dilution or propidium iodide staining, for example. Measuring the increase or decrease in induction of complement-dependent cytotoxicity and / or antibody-dependent cell-mediated cytotoxicity. Increased activity indicates immunostimulatory activity. A suitable increase in activity is outlined below.

  In one embodiment, T cell activation is, for example, by direct killing of a target cell, such as a cancer cell, or by cytokine secretion, or by proliferation, or activation such as, for example, CD137, CD107a, PD1, etc. Measured by changes in marker expression. For T cells, proliferation, cell surface markers of activation (eg, CD25, CD69, CD137, PD1), cytotoxicity (ability to kill target cells), and cytokine production (eg, IL-2, IL-4) , IL-6, IFNγ, TNF-a, IL-10, IL-17A) can be an indicator of immunomodulation consistent with enhanced killing of cancer cells.

  In one embodiment, NK cell activation is measured, for example, by direct killing of a target cell, eg, a cancer cell, or by cytokine secretion, or by a change in the expression of an activation marker, eg, CD107a. . For NK cells, proliferation, cytotoxicity (ability to kill target cells and increase CD107a, granzyme, and perforin expression), cytokine production (eg, IFNγ and TNF), and cell surface receptor expression (eg, CD25) ) Increase can be an indicator of immune regulation consistent with enhanced killing of cancer cells.

  In one embodiment, T cell activation of γδ is measured, for example, by cytokine secretion, or by proliferation, or by altered expression of an activation marker.

  In one embodiment, Th1 cell activation is measured, for example, by cytokine secretion or by altered expression of activation markers.

  A suitable increase (or decrease) in activity or response, as appropriate as outlined above, is relative to the signal in either a reference sample or a control sample, eg, a test sample that does not contain an anti-PVRIG antibody of the invention. , At least 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, 90%, 95%, or 98-99% percent increase. Similarly, an increase of at least 1 fold, 2 fold, 3 fold, 4 fold or 5 fold is effective compared to a reference subject or control sample.

XI. Treatment Once manufactured, the composition of the invention generally inhibits suppression of T cell activation, typically by linking a bispecific immunomodulatory antibody of the invention (eg, T cells are no longer suppressed). Is used for several oncology applications by treating cancer.

  Accordingly, the heterodimeric compositions of the present invention are used for the treatment of these cancers.

XII. Antibody Compositions for In Vivo Administration An antibody formulation for use in accordance with the present invention may comprise an antibody of desired purity with any pharmaceutically acceptable carrier, excipient, or stabilizer (Remington's Pharmaceutical Sciences 16th prepared for storage in the form of a lyophilized formulation or an aqueous solution by mixing with (as generally outlined in edition, Osol, A. Ed. [1980]). Acceptable carriers, buffers, excipients, or stabilizers are nontoxic to recipients at the dosages and concentrations used and buffers such as phosphate, citrate, and other organic acids; Antioxidants including ascorbic acid and methionine; preservatives (octadecyldimethylbenzylammonium chloride; hexamethonium chloride; benzalkonium chloride, benzethonium chloride; phenol, butyl, or benzyl alcohol; alkyl parabens such as methyl or propylparaben; catechol Resorcinol; cyclohexanol; 3-pentanol; and m-cresol); low molecular weight (less than about 10 residues) polypeptide; protein such as serum albumin, gelatin, or immunoglobulin; hydrophilic polymer such as polyvinylpyrrolidone; glycine The Amino acids such as tamin, asparagine, histidine, arginine, lysine; monosaccharides, disaccharides and other carbohydrates including glucose, mannose, or dextrin; chelating agents such as EDTA; sugars such as sucrose, mannitol, trehalose or sorbitol; sodium Salt-forming counterions such as: metal complexes (eg, Zn-protein complexes); and / or nonionic surfactants such as TWEEN ™, PLURONICS ™ or polyethylene glycol (PEG).

Mode of Administration The antibodies and chemotherapeutic agents of the invention are administered to a subject according to known methods such as intravenous administration as a bolus or continuous infusion over a period of time.

Treatment Methods In the methods of the present invention, treatment is used to provide a positive therapeutic response with respect to a disease or condition. A “positive therapeutic response” is intended to ameliorate a disease or condition and / or ameliorate symptoms associated with the disease or condition. For example, a positive therapeutic response may refer to one or more of the following improvements in the disease: (1) decreased number of tumor cells, (2) increased tumor cell death, (3) tumor cell survival. Inhibition of (5) inhibition of tumor growth (ie, some slowing, preferably cessation), (6) increased patient survival, and (7) some from one or more symptoms associated with the disease or condition release.

  A positive therapeutic response in any given disease or condition can be determined by standardized response criteria specific to that disease or condition. Tumor response includes tumors including magnetic resonance imaging (MRI) scans, radiography, computed tomography (CT) scans, bone scans, endoscopy, and bone marrow aspiration (BMA) and counting of circulating tumor cells Screening techniques such as biopsy sampling can be used to assess changes in tumor morphology (ie, whole body tumor tissue volume, tumor size, etc.).

  In addition to these positive therapeutic responses, the subject being treated may experience the beneficial effects of ameliorating symptoms associated with the disease.

  Treatment according to the present invention includes a “therapeutically effective amount” of the pharmaceutical agent used. “Therapeutically effective amount” refers to an amount that is effective at the doses and for periods required to achieve the desired therapeutic result.

  A therapeutically effective amount may vary depending on factors such as the disease state, age, sex, and weight of the individual, and the ability of the agent to elicit a desired response in the individual. A therapeutically effective amount is also an amount that provides a therapeutically beneficial effect over any toxic or deleterious effect of the antibody or antibody portion.

  A “therapeutically effective amount” for tumor therapy can also be measured by its ability to stabilize disease progression. The ability of a compound to inhibit cancer may be evaluated in an animal model system that predicts efficacy in human tumors.

  Alternatively, this property of the composition may be assessed by examining the ability of the compound to inhibit cell proliferation or induce apoptosis by in vitro assays known to those skilled in the art. A therapeutically effective amount of a therapeutic compound can reduce tumor size or otherwise reduce a subject's symptoms. One of ordinary skill in the art will be able to determine such amounts based on factors such as the size of the subject, the severity of the subject's symptoms, and the particular composition or route of administration selected.

  Dosage regimens are adjusted to provide the optimum desired response (eg, a therapeutic response). For example, a single bolus may be administered, several divided doses may be administered over time, or the dose may be reduced or increased proportionally as indicated by the urgency of the treatment situation Good. Parenteral compositions can be formulated in dosage unit form for ease of administration and uniformity of dosage. A unit dosage form as used herein refers to a physically discrete unit suitable as a single dose for the subject being treated, each unit as desired in connection with the required pharmaceutical carrier. A predetermined amount of active compound calculated to produce a therapeutic effect of

  The specification of the unit dosage form of the present invention is the field of formulating such active compounds for (a) the unique characteristics of the active compound and the specific therapeutic effect to be achieved, and (b) the treatment of susceptibility in an individual. Is affected by and depends directly on the restrictions that follow.

  The effective dosage and dosing regimen of the bispecific antibody used in the present invention will depend on the disease or condition being treated and can be determined by one skilled in the art.

  An exemplary and non-limiting range for a therapeutically effective amount of a bispecific antibody used in the present invention is about 0.1 to 100 mg / kg.

  All references are expressly incorporated herein by reference in their entirety.

  While particular embodiments of the present invention have been described above for purposes of illustration, those skilled in the art will recognize that numerous changes in detail may be made without departing from the invention as set forth in the appended claims. It will be understood.

VIII. EXAMPLES Examples are provided below to illustrate the present invention. These examples are not meant to limit the invention to any particular application or theory of operation. For all the constant region positions discussed in the present invention, the numbering is performed by Kabat (Kabat et al., 1991, Sequences of Proteins of Immunological Institute, 5th Ed., United States Public Health, Public Health Public Health. Follow the same EU index as fully embedded). One skilled in the art of antibodies will understand that this rule consists of non-contiguous numbering in specific regions of the immunoglobulin sequence, allowing a standard reference to a conserved position in the immunoglobulin family. . Thus, the position of any given immunoglobulin as defined by the EU index does not necessarily correspond to its contiguous sequence.

  General and specific science and technology are outlined in US Patent Application Publication Nos. 2015/0307629, 2014/0288275, and WO 2014/145806, all of which relate specifically to the technology outlined therein. , Explicitly incorporated by reference in its entirety.

A. Example 1: TIL from multiple cancer types co-express PD-1 and T cell costimulatory receptors To investigate the potential relevance between PD-1 and various T cell costimulatory receptors In addition, RNA sequencing data from The Cancer Genome Atlas Project (TCGA) was used for analysis. V2 RSEM data was downloaded from FireBrowse (http://firebrowse.org/). Analysis was performed using R by routine routine. The correlation between the expression of PD-1 and the eight costimulatory receptors is shown in FIG. 1 along with the calculated R2 value (square of the Pearson correlation coefficient). Data show that PD-1 and some costimulatory receptors are bladder cancer, breast cancer, colon cancer, head and neck cancer, kidney cancer, lung adenoma, lung squamous cell carcinoma, ovarian cancer, pancreatic cancer, prostate cancer, and black Co-expressed in cancers including tumors. In particular, ICOS expression on TIL correlates better with PD-1 expression than some other costs.

B. Example 2: Immune checkpoint antigen linking domain 1.2A: Anti-PD-1 ABD
An example of an antibody linking to PD-1 was generated in a bivalent IgG1 format with an exemplary sequence having the E233P / L234V / L235A / G236del / S267K substitution shown in FIG. DNA encoding the variable region was generated by gene synthesis and inserted into the mammalian expression vector pTT5 by the Gibson Assembly method. The heavy chain VH gene was inserted via Gibson assembly into pTT5 encoding the human IgG1 constant region with the above substitutions. The light chain VL gene is human C? It was inserted into pTT5 encoding the constant region. DNA was transfected into HEK293E cells for expression. Additional PD-1 ABDs (including those derived from the above antibodies) were formatted as Fab and scFv for use in costim / checkpoint bispecific antibodies, with exemplary sequences shown in FIG. 11 and the sequence listing, respectively. Shown in

2.2B: Anti-CTLA-4 ABD
Antibodies linking to CTLA-4 were generated in a bivalent IgG1 format with E233P / L234V / L235A / G236del / S267K substitution. DNA encoding the variable region was generated by gene synthesis and inserted into the mammalian expression vector pTT5 by the Gibson Assembly method. The heavy chain VH gene was inserted via Gibson assembly into pTT5 encoding the human IgG1 constant region with the above substitutions. The light chain VL gene is human C? It was inserted into pTT5 encoding the constant region. DNA was transfected into HEK293E cells for expression. Additional CTLA-4 ABD (including those derived from the above antibodies) was formatted as scFv for use in a costim / checkpoint bispecific antibody, an exemplary sequence of which is shown in FIG. 12 and the sequence listing.

3.2C: Anti-LAG-3 ABD
An example of an antibody linking to LAG-3 was generated in a bivalent IgG1 format with E233P / L234V / L235A / G236del / S267K substitution. DNA encoding the variable region was generated by gene synthesis and inserted into the mammalian expression vector pTT5 by the Gibson Assembly method. The heavy chain VH gene was inserted via Gibson assembly into pTT5 encoding the human IgG1 constant region with the above substitutions. The light chain VL gene is human C? It was inserted into pTT5 encoding the constant region. DNA was transfected into HEK293E cells for expression. Additional LAG-3 ABD (including those derived from the above antibodies) was formatted as a Fab for use in a costim / checkpoint bispecific antibody, an exemplary sequence of which is shown in FIG. 13 and the sequence listing.

4.2D: Anti-TIM-3 ABD
An example of an antibody linking to TIM-3 was generated in a divalent IgG1 format with an exemplary sequence having the E233P / L234V / L235A / G236del / S267K substitution shown in FIG. DNA encoding the variable region was generated by gene synthesis and inserted into the mammalian expression vector pTT5 by the Gibson Assembly method. The heavy chain VH gene was inserted via Gibson assembly into pTT5 encoding the human IgG1 constant region with the above substitutions. The light chain VL gene is human C? It was inserted into pTT5 encoding the constant region. DNA was transfected into HEK293E cells for expression. The antibody described above was formatted as a Fab for use in a costim / checkpoint bispecific antibody.

5.2E: Anti-PD-L1 ABD
Prototype antibodies linking to PD-L1 were generated in a bivalent IgG1 format with an exemplary sequence having the E233P / L234V / L235A / G236del / S267K substitution shown in FIG. DNA encoding the variable region was generated by gene synthesis and inserted into the mammalian expression vector pTT5 by the Gibson Assembly method. The heavy chain VH gene was inserted via Gibson assembly into pTT5 encoding the human IgG1 constant region with the above substitutions. The light chain VL gene is human C? It was inserted into pTT5 encoding the constant region. DNA was transfected into HEK293E cells for expression. The antibodies described above were formatted as Fab and scFv for use in costim / checkpoint bispecific antibodies.

C. Example 3: Costimulatory Receptor Antigen Binding Domain Prototype costimulatory receptor antibodies linking to ICOS, GITR, OX40, and 4-1BB are in a bivalent IgG1 format with E233P / L234V / L235A / G236del / S267K substitution. And their sequences are shown in FIGS. DNA encoding the variable region was generated by gene synthesis and inserted into the mammalian expression vector pTT5 by the Gibson Assembly method. The heavy chain VH gene was inserted via Gibson assembly into pTT5 encoding the human IgG1 constant region with the above substitutions. The light chain VL gene is human C? It was inserted into pTT5 encoding the constant region. DNA was transfected into HEK293E cells for expression. The above-described antibodies were formatted as Fab or scFv for use in costim / checkpoint bispecific antibodies.

D. Example 4: Manipulation of anti-ICOS ABD for stability and affinity The parental variable region of the anti-ICOS antibody (XENP16435; shown in FIG. 19) is for use as a component in the ICOS × checkpoint bispecific antibody. Was operated on. Library of Fv mutations engineered to have optimal affinity and stability can be site-directed mutagenesis (QuikChange, Stratagene, Cedar Creek, Tx) or further gene synthesis and in Fab-His and scFv-His formats It was constructed by subcloning and prepared as follows.

1.4A: Anti-ICOS Fab
The amino acid sequence of the mutant anti-ICOS Fab is listed in FIG. 20 (the polyhistidine (His6) tag is removed from the C-terminus of the Fab heavy chain). DNA encoding the two strands required for Fab expression was generated by gene synthesis and subcloned into the expression vector pTT5 using standard molecular biology techniques. The Fab heavy chain contained a C-terminal polyhistidine tag. DNA was transfected into HEK293E cells for expression and the resulting protein was purified from the supernatant using Ni-NTA chromatography. The resulting anti-ICOS Fab was analyzed for stability and affinity.

  Differential scanning fluorescence analysis (DSF) experiments were performed using a Bio-Rad CFX Connect real-time PCR detection system. The protein was mixed with SYPRO orange fluorescent dye and diluted to 0.2 mg / mL with PBS. The final concentration of SYPRO orange was 10 times. Following the first 10 minute incubation period at 25 ° C., the protein was heated from 25 ° C. to 95 ° C. at a heating rate of 1 ° C./min. Fluorescence measurements were performed every 30 seconds. Melting temperature (Tm) was calculated using instrument software. The results are shown in FIG.

  A series of affinity screens of anti-ICOS Fabs against human ICOS were performed using Octet, a method based on BioLayer Interferometry (BLI). Octet's experimental steps generally included: immobilization (capture of ligand or test sample on biosensor), association (ligand or well into corresponding test sample or well containing serial dilutions of ligand or Immersion of the biosensor coated with the test sample), dissociation to determine the monovalent affinity of the test sample (returning the biosensor to the well containing the buffer). Specifically, an anti-mouse Fc (AMC) biosensor was used to capture ICOS mouse IgG2a Fc fusion and soak in multiple concentrations of test samples. The resulting equilibrium dissociation constant (KD), association rate (ka), and dissociation rate (kd) are presented in FIGS. Ligation affinity and dynamic rate constants were obtained by analyzing processed data using a 1: 1 ligation model using ForteBio Octet Data Analysis software (ForteBio). Data from two separate experiments are shown in Figures 22A-B. In further experiments, a streptavidin (SA) biosensor was used to capture ICOS-TEV-Fc-Avi and soak in the test sample. The data is shown in FIG. Several mutant anti-ICOS Fabs, including XENP22780, XENP22778, XENP22783, and XENP22784, had improved stability while maintaining similar affinity properties as the Fab containing the parental variable region (XENP22050).

2.4B: Anti-ICOS scFv
The amino acid sequence of anti-ICOS scFv is listed in FIG. 24 (the polyhistidine (His6) tag has been removed from the C-terminus of scFv). DNA encoding scFv was generated by gene synthesis and subcloned into the expression vector pTT5 using standard molecular biology techniques. The scFv contained a C-terminal polyhistidine tag. DNA was transfected into HEK293E cells for expression and the resulting protein was purified from the supernatant using Ni-NTA chromatography. The resulting anti-ICOS scFv was analyzed for stability in DSF experiments as described above. The data is shown in FIG.

E. Example 5: Anti-ICOS x Anti-PD-1 Bispecific Antibody 1.5A: Prototype costim / checkpoint bottle opener A schematic diagram of a costim / checkpoint bispecific antibody in a bottle opener format is shown in FIG. 2A. Co-stimulatory receptor-linked Fab arms based on prototype anti-GITR mAb (XENP16438), anti-OX40 mAb (XENP16437), anti-4-1BB mAb (XENP14410), and anti-ICOS mAb (XENP16435), and exemplary anti-PD-1 Prototype bottle openers with scFv (XENP19692) arms were produced to investigate their effects on cytokine secretion in the SEB-stimulated PBMC assay. The sequence is shown in FIG. DNA encoding the three strands required for bottle opener expression was generated by gene synthesis and subcloned into the expression vector pTT5 using standard molecular biology techniques. HEK293E cells were DNA transfected for expression and the resulting protein was purified using standard techniques.

a. 5A (a): Prototype anti-ICOS × anti-PD-1 bottle opener enhances cytokine secretion Staphylococcal enterotoxin B (SEB) is achieved by activation through the T cell receptor (TCR) It is a superantigen that causes T cell activation and proliferation. Stimulation of human PBMC with SEB is a common method for assaying T cell activation and proliferation. PBMC were simulated with 100 ng / mL SEB for 2 days. Cells were washed twice and restimulated with 100 ng / mL SEB in combination with the indicated test sample at 20 μg / mL. A first bivalent anti-PD-1 antibody based on nivolumab (XENP16432), a second bivalent anti-PD-1 mAb (XENP19686), and a bivalent anti-RSV mAb (XENP15074) were used as controls. After 24 hours of treatment, the supernatant was assayed for IL-2. The data shown in FIG. 27 shows that each costim / checkpoint bottle opener enhanced cytokine secretion compared to the bivalent anti-RSV antibody. Surprisingly, induction of cytokine secretion by XENP 22730 is far superior to cytokine secretion by the bivalent anti-PD-1 antibody alone as well as other prototype costim / checkpoint bottle openers, with the addition of ICOS linkage being the cytokine We show that production is enhanced and that ICOS is a better PD-1 partner than other costimulatory receptors for bispecific antibodies.

  In another experiment, PBMC were stimulated with 0.01 μg / mL SEB for 3 days with the indicated test sample at 20 μg / mL. As a control, test samples were also incubated with untreated (non-SEB stimulated) PBMC. Three days after treatment, the supernatants were evaluated for IL-2 secretion as an indicator of T cell activation (shown in FIG. 28). The data show that neither anti-PD-1 bivalent antibody nor anti-ICOS × anti-PD-1 bottle opener stimulates IL-2 secretion in naive cells. In addition, the data similarly indicate that XENP20896 enhances IL-2 secretion over the bivalent anti-PD-1 antibody alone, and that XENP20896 is a bivalent combination (XENP16432 and XENP16435), as well as a one-arm combination (XENP20111 and XILP20266) shows enhanced IL-2 secretion.

  Costimulatory receptors such as ICOS have previously been found to induce cytokine production only after cross-linking with bivalent antibodies or multimerizing ligands (Viera et al. 2004, Sannamed et al. 2015). Given the cross-linking mechanism, it is surprising that monovalent ICOS ligation by a single arm in a costim x checkpoint blocking bispecific antibody could enhance cytokine production. Furthermore, it is noteworthy that bispecific antibodies were able to enhance cytokine secretion over bivalent anti-PD-1 mAbs (XENP16432 + XENP16435) combined with bivalent anti-ICOS mAbs.

b. 5A (b): PD-1 and ICOS double positive cells are selectively occupied by prototype anti-ICOS × anti-PD-1 bottle openers Non-tumor response expressing only immune checkpoint receptors or costimulatory receptors Selective targeting of tumor-reactive TIL co-expressing immune checkpoint receptors (eg PD-1) and costimulatory receptors as shown in Example 1 to sex T cells while increasing anti-tumor activity On the other hand, local toxicity is avoided (as shown in FIG. 29).

  The SEB-stimulated PBMC assay was used to examine the binding of anti-ICOS × anti-PD-1 bottle opener to T cells. PBMC were stimulated with 100 ng / mL SEB (staphylococcal enterotoxin B) for 3 days, after which PBMC were treated with the indicated test samples at 4 ° C. for 30 minutes. PBMC were then incubated with APC-labeled 1-arm anti-ICOS antibody and FITC-labeled 1-arm anti-PD-1 antibody for 30 minutes at 4 ° C. FIGS. 30 and 31 show receptor occupancy of prototype anti-ICOS × anti-PD-1 bottle opener (XENP20896), 1-arm anti-ICOS antibody (XENP20266), and 1-arm anti-PD-1 antibody (XENP20111).

  Data show that double positive cells (expressing both PD-1 and ICOS) are selectively occupied by anti-ICOS × anti-PD-1 bottle opener (XENP20896) as shown in FIG. , Indicating that monovalent ICOS and PD-1 linkages are useful for selective targeting. Furthermore, anti-ICOS × anti-PD-1 bottle openers (eg, XENP20896) are more potent in double positive cells than monovalent monospecific 1-arm anti-PD-1 and anti-ICOS, as shown in FIG. 31A. Link.

c. 5A (c): Prototype anti-ICOS x anti-PD-1 bottle opener promotes engraftment in GVHD mouse test Prototype anti-ICOS x anti-PD-1 bottle opener is tested in NSG (NOD-SCID-gamma) immunodeficient mice Evaluated in a performed graft versus host disease (GVHD) model. Mice were transplanted with human PBMC. When NSG mice are injected with human PBMC, they develop an autoimmune response against human PBMC. Treatment of NSG mice injected with human PBMC with an immune checkpoint antibody (eg, anti-PD-1) promotes engraftment. Thus, increased engraftment indicates the effectiveness of the antibody.

  Ten million human PBMCs were transplanted into NSG mice via IV-OSP on day 0, followed by administration of the test samples indicated on days 1, 8, 15, and 22. The number of human CD45 + cells was measured on day 14 as an indicator of disease.

  The data shown in FIG. A shows that anti-ICOS × anti-PD-1 bottle opener enhances CD45 + cell proliferation in human PBMC transplanted NSG mice compared to controls (PBS and PBS + PBMC). Furthermore, using the antibodies of the present invention, the enhancement is greater than that seen with bivalent anti-PD-1 antibodies.

2.5B: Production of mutant anti-ICOS with optimized anti-ICOS Fab arm x anti-PD-1 bottle opener Mutant anti-ICOS with anti-ICOS Fab engineered as described in Example 4A Anti-PD-1 bottle openers were produced generally as described above. The amino acid sequence of the mutant anti-ICOS × anti-PD-1 bottle opener is listed in FIG. The amino acid sequence of mutant anti-ICOS × anti-PD-1 having an affinity enhancing substitution of FcRn pH 6.0 is listed in FIG.

  The resulting anti-ICOS × anti-PD-1 bispecific antibody was analyzed for affinity for human and cynomolgus monkey ICOS using Octet as generally described above. In the first set of experiments, an AMC biosensor was used to capture mouse IgG2a Fc fusion of ICOS and soak in multiple concentrations of test samples (data shown in FIG. 34). In further experiments, a streptavidin (SA or SAX) biosensor was used to capture human and cynomolgus ICOS biotinylated human and cynomolgus IgG1 Fc fusions and soaked in multiple concentrations of test samples (data shown in FIG. 35). To show).

3.5C: T cell surface ligation of mutant anti-ICOS × anti-PD-1 bottle opener Anti-ICOS × anti-PD-1 bispecific ligation to T cells was measured by SEB-stimulated PBMC assay. Human PBMC were stimulated with 100 ng / mL SEB for 3 days. PBMC were then treated with the indicated test article samples and incubated at 4 ° C. for 30 minutes. After treatment, cells were incubated with FITC-labeled anti-CD3 antibody and APC-labeled anti-human IgG Fc secondary antibody. The MFI on CD3 + cells is shown in FIGS.

4.5D: Receptor occupancy of mutant anti-ICOS x anti-PD-1 bottle opener on T cells Receptor occupancy of mutant anti-ICOS x anti-PD-1 bottle opener on T cells was measured by SEB-stimulated PBMC assay. . PBMC were stimulated with 100 ng / mL SEB (staphylococcal enterotoxin B) for 3 days, after which PBMC were treated with the indicated test samples at 4 ° C. for 30 minutes. PBMC were then incubated with APC-labeled 1-arm anti-ICOS antibody and FITC-labeled 1-arm anti-PD-1 antibody for 30 minutes at 4 ° C. FIG. 37 shows the mutant anti-ICOS × anti-PD-1 bispecific antibody, the corresponding 1-arm anti-ICOS antibody and 1-arm anti-PD-1 antibody (XENP20111) in PD-1 and ICOS double positive T cells. Shows receptor occupancy. Consistent with the prototype antibody examined in Example 1A, each of the bottle openers binds more strongly to double positive cells than the monovalent monospecific 1-arm anti-PD-1 antibody and 1-arm anti-ICOS antibody.

5.5E: In vitro activity of mutant anti-ICOS x anti-PD-1 bottle opener in cytokine release assay. Human PBMC were stimulated with 100 ng / mL SEB for 2 days. Cells were washed and restimulated with 100 ng / mL SEB in combination with the indicated test sample at 20 μg / mL. After 24 hours of treatment, cells were treated with IL-2 (FIG. 38A) and IFN? Assayed for. (FIG. 38B). Data show that anti-ICOS × anti-PD-1 bottle opener is bivalent anti-PD-1 antibody (XENP16432) alone, bivalent anti-ICOS antibody (XENP16435) alone, or bivalent anti-PD-1 antibody and bivalent anti-ICOS It shows that significantly more cytokine release was stimulated than the antibody combination.

  In further experiments, human PBMC were stimulated with 100 ng / mL SEB for 2 days. Cells were washed and restimulated with 100 ng / mL SEB in combination with the indicated test sample at 20 μg / mL. After 24 hours of treatment, the cells were treated with IL-2 (FIG. 39A) and IFN? Assayed. (FIG. 39B).

6.5F: In Vivo Activity of Mutant Anti-ICOS x Anti-PD-1 Bottle Opener in GVHD Mouse Study In the first study, 10 million human PBMCs were transplanted on day 0 via IV-OSP to NSG mice. Subsequently, the test sample indicated on day 1 was administered. On the 7th, 11th and 14th days, IFN? Levels and numbers of human CD45 +, CD8 + T cells and CD4 + T cells were measured. 40A-B show the IFN on the 7th and 11th days, respectively? Indicates the level. 41A-B show the number of CD45 + cells on day 11 and day 14, respectively. 42A-B show the number of CD8 + T cells and CD4 + T cells on day 14, respectively. FIG. 43 shows T cell proliferation and IFN? Figure 7 shows the change in body weight of mice up to day 14 resulting from GVHD deterioration due to production.

  In a further study with additional mutant anti-ICOS × anti-PD-1 bottle opener, 10 million human PBMCs were transplanted via IV-OSP into NSG mice on day 0, followed by day 1 Test samples indicated at the indicated concentrations were administered. On days 7 and 14, IFNγ levels and human CD45 +, CD8 + T cells and CD4 + T cell numbers were measured. 44A-B show the IFN on day 7 and 14 respectively? Indicates the level. FIG. 45 shows the number of CD45 + cells on day 14. FIGS. 46A-C show the CD8 + T cell count, CD4 + T cell count, and CD8 + / CD4 + ratio on day 14, respectively. Mouse body weights were also measured on day 12 and day 15 and are shown as percentages of initial body weight in Figures 47A-B, respectively.

  The figure shows that anti-ICOS × anti-PD-1 bottle opener enhances engraftment (as shown by proliferation of CD45 + cells, CD8 + T cells and CD4 + T cells and weight loss of mice). The observed activity correlates with the in vitro potency of each mutation. In addition, several bottle openers including XENP20896, XENP22744, XENP23092, XENP22730, XENP23104, XENP22974, and XENP23411 enhance engraftment far more than the control bivalent anti-PD-1 antibody (XENP16432).

7.5G: Further anti-ICOS ABD functions with anti-ICOS × anti-PD-1 bottle opener Additional anti-ICOS × anti-PD comprising anti-ICOS Fabs based on other anti-ICOS ABDs as described in Example 3 generally as described above. -1 bottle opener was produced. An arrangement for a further bottle opener is shown in FIG.

  In the first experiment, PBMC were stimulated with 100 ng / mL SEB for 2 days. Cells were washed twice and restimulated with 100 ng / mL SEB in combination with 20 μg / mL indicated test sample (PBS as a control). After 24 hours of treatment, the supernatants were assayed for IL-2 concentration as shown in FIG. In a second experiment, PBMC were stimulated with 10 ng / mL SEB and treated with 20 μg / mL of the indicated test sample for 3 days (bivalent anti-PD-1XENP16432 as a control). The supernatant was collected and assayed for IL-2. FIG. 50 shows the induction factor of IL-2 against bivalent anti-RSV mAb.

  Consistent with the data in Example 5A (a), XENP20896 stimulated IL-2 secretion. In particular, additional bottle openers containing alternative anti-ICOS ABDs can also stimulate IL-2 secretion, and enhancement of cytokine secretion by anti-ICOS × anti-PD-1 bottle openers can be achieved using the parent described in Example 4. To demonstrate that it is not unique to ABD derived from the anti-ICOS ABD.

8.5H: Anti-ICOS x anti-PD-1 bottle opener shows clear signs of ICOS a. 5H (a): anti-ICOS × anti-PD-1 bispecific antibody induces AKT phosphorylation ICOS ligation induces AKT phosphorylation in activated T cells (Fos, C et al. 2008) and As such, AKT phosphorylation will be an indicator of ICOS agonism by the anti-ICOS × anti-PD-1 bispecific antibody of the present invention.

  Human PBMC were stimulated with 100 ng / mL SEB for 2 days. After stimulation, CD3 + cells were isolated by negative selection using EasySep ™ human T cell enrichment kit (STEMCELL Technologies, Vancouver, Canada) and then shown in combination with plate-linked anti-CD3 antibody (OKT3, 500 ng / mL) The test sample was treated. Cells were lysed 30 minutes after treatment and assayed for total AKT and phosphorylated AKT (Ser473) by multiple phosphorylated protein assay on MULTI-SPOT 384 well spot plates (Meso Scale Discovery, Rockville, Md.). Data is shown in FIG. 51 as the percentage of AKT phosphorylated after treatment.

  The data does not show an increase in AKT phosphorylation after treatment with the negative control bivalent anti-PD-1 mAb (XENP16432). Both the bivalent anti-ICOS mAb alone (XENP16435) and XENP16435 in combination with XENP16432 increased AKT phosphorylation, demonstrating ICOS ligation and agonism by XENP16435. Surprisingly, despite only monovalent linkage of ICOS, treatment with anti-ICOS × anti-PD-1 bispecific antibody resulted in more AKT phosphorus than treatment with XENP16435 alone and XENP16435 and XENP16432. Induces oxidation. Positive AKT phosphorylation data demonstrates a clear indication of ICOS activity with the bispecific antibody despite the monovalent linkage of ICOS.

b. 5H (b): activated T helper cell-related gene upregulated by anti-ICOS × anti-PD-1 bottle opener Guedan et al. (2014) describes gene expression profiles of activated helper T cells (ie Th1, Th2, and Th17) and regulatory T cells (Treg) after activation of CAR-Ts based on the ICOS signaling domain. . Are they IL-17A, IL22, IFN? We found that genes associated with activated T helper cells such as TNF, TNF, and IL-13 were upregulated, whereas Treg-related genes such as TGFβ1, SMAD3, and FOXP3 were unchanged or downregulated .

  In order to determine whether ligation of T cells with the anti-ICOS × anti-PD-1 bispecific antibody of the present invention resulted in similar symptoms, we used Nanostring technology. PBMC were stimulated with 100 ng / mL SEB for 2 days. Cells were washed twice and restimulated with 100 ng / mL SEB and treated with the indicated test samples for 24 hours. RNA was extracted from the cells and assayed by nCounter® PanCancer Immunoprofiling Panel (NanoString Technologies, Seattle, Wash.), Which assayed 770 target genes including immune responses. FIG. 52 shows the mean fold induction in the expression of several Th-related and Treg-related genes for bivalent anti-RSV antibodies. 53A-F show IL-17A, IL-17F, IL-22, IL-10, IL-9, and IFN, respectively? Shows the fold induction of gene expression by the indicated test sample over that induced by the bivalent anti-RSV mAb.

  As shown in FIGS. 52-53 and Gueden et al. Consistent with observations by IL-17A, IL-17F, IL-22, IL-9, and IFN? Th-related genes related to ICOS signaling such as are up-regulated after treatment with bivalent anti-ICOS mAb and combinations of anti-ICOS and anti-PD-1 mAbs. In particular, the expression of Th-related genes associated with ICOS signaling is further upregulated after treatment with anti-ICOS × anti-PD-1 antibody, which is also evident by anti-ICOS costimulatory signs and anti-bispecific antibodies. Shows a dramatic synergy of ICOS and anti-PD-1 ligation.

9.5I: Alternative format anti-ICOS x anti-PD-1 bispecific antibody The effect of anti-ICOS x anti-PD-1 bottle opener is specific to the bottle opener format or anti-ICOS x anti-PD An alternative format costim / checkpoint bispecific antibody was produced to see if it was widely applicable to -1 bispecific antibodies.

a. 5I (a): anti-PD-1 × anti-ICOS bottle opener Produced using anti-PD-1 mAbs as described in Example 2A (eg, XENP16432 and XENP29120) and DNA encoding anti-ICOS scFv as described in Example 4B An anti-PD-1 × anti-ICOS bottle opener with anti-PD-1 Fab was generated. A bottle opener was produced as generally described in Example 5A. An exemplary anti-PD-1 × anti-ICOS bottle opener array is shown in FIG.

b. 5I (b): center-scFv
A schematic of the center-scFv format is shown in FIGS. DNA encoding the anti-ICOS Fab-Fc heavy chain was generated by standard subcloning into the expression vector pTT5 using the DNA encoding the anti-ICOS Fab described in Example 4A. The anti-ICOS Fab described in Example 4A and the DNA encoding the anti-PD-1 scFv described in Example 2A were used by a combination of gene synthesis and subcloning into the standard expression vector pTT5 to produce anti-ICOS-anti-PD. A DNA encoding the -1 Fab-scFv-Fc heavy chain was generated. DNA was transfected into HEK293E cells for expression. The sequence of an exemplary anti-ICOS × anti-PD-1 middle-scFv antibody is shown in FIG.

c. 5I (c): middle-scFv2
A schematic of the center-scFv2 format is shown in FIG. 55C. Using a combination of gene synthesis and subcloning into the standard expression vector pTT5 using the anti-ICOS Fab described in Example 4A and the DNA encoding the anti-PD-1 scFv described in Example 2A, anti-ICOS-anti-PD A DNA encoding the -1 Fab-scFv-Fc heavy chain was generated. DNA was transfected into HEK293E cells for expression. The sequence of an exemplary anti-ICOS × anti-PD-1 middle-scFv2 antibody is shown in FIG.

d. 5I (d): Bispecific mAb
A schematic diagram of the bispecific mAb format is shown as FIG. 2K. DNA encoding anti-ICOS heavy and light chains is generated based on DNA encoding anti-ICOS Fab as described in the Examples, and DNA encoding anti-PD-1 heavy and light chains is described in Example 2A. Produced based on DNA encoding the described anti-PD-1 mAb. The heavy chain VH gene was inserted via Gibson assembly into pTT5 encoding a human IgG1 constant region with E233P / L234V / L235A / G236del / S267K substitution. The light chain VL gene is human C? It was inserted into pTT5 encoding the constant region.

  DNA was transfected into HEK293E cells for expression as two separate antibodies and purified by protein A affinity (GE Healthcare). These antibodies contain a heterodimerization-skew substitution at the CH3: CH3 interface. 2-mercaptoethylamine-HCl (2-MEA) was used to induce controlled reduction of interchain disulfide linkages in the two parent IgGs. 2-MEA was then removed, resulting in re-oxidation of the interchain disulfide linkage, allowing re-ligation of the HC-LC pair (driven by a heterodimerization mutation). Finally, the trispecific antibody was purified by cation exchange chromatography. Such methods are common in the art (see, eg, Labrijn (2013) PNAS110 (13): 5145-50 or Strop et al. (2012) J Mol Biol 420 (3): 204-19). Wanna) Other design and purification methods known in the art can be used to produce bispecific mAbs such as common light chain antibodies (Merchant (1998) Nat Biotechnol 16 (7): 677-81) or heterobiotics. Production of the monomeric Fab domain (Lewis et al. (2014) Nat Biotechnol 32 (2): 191-8) can be promoted. The sequence of an exemplary anti-PD-1 × anti-ICOS bispecific mAb is shown in FIG.

e. 5I (e): DVD-IgG
A schematic diagram of the DVD-IgG format is shown in FIG. 55E. Anti-PD-1-anti-ICOS VH-VH-CH1-Fc heavy chain and VL-VL-C? The DNA encoding the light chain can be obtained by standard subcloning into the gene synthesis and expression vector pTT5 using DNA encoding the anti-ICOS Fab described in Example 4A and the anti-PD-1 scFv described in Example 2A. Generated by combination. DNA was transfected into HEK293E cells for expression. An exemplary anti-ICOS × anti-PD-1 DVD-IgG sequence is shown in FIG.

f. 5I (f): Trident A schematic diagram of the trident format is shown in FIG. 55F. The DNA encoding the anti-PD-1 VL-VH-Fc heavy chain and VL-VH light chain is the DNA encoding the anti-ICOS Fab described in Example 4A and the anti-PD-1 scFv described in Example 2A. Generated by a combination of gene synthesis and standard subcloning into the expression vector pTT5. DNA was transfected into HEK293E cells for expression. An exemplary anti-ICOS × anti-PD-1 trident array is shown in FIG.

g. 5I (g): An alternative format of anti-ICOS × anti-PD-1 bispecific antibody enhances cytokine secretion Human PBMC were stimulated with 200 ng / mL SEB for 2 days. Cells were washed twice and restimulated with 100 ng / mL SEB in combination with the indicated test samples at the indicated concentrations. After 24 hours of treatment, the supernatant was assayed for IL-2. The data is shown in FIG. 62, showing that each of the alternative format bispecific antibodies enhances IL-2 secretion compared to the bivalent anti-RSV (XENP15074) control. In particular, the majority of these alternative format antibodies enhance IL-2 secretion compared to bivalent anti-PD-1 antibodies.

F. Example 6: Costim / Checkpoint Bispecific Antibody Targeting Different Immune Checkpoint Antigen 1.6A: Anti-CTLA-4x Anti-ICOS
An anti-CTLA-4 × anti-ICOS bottle opener was produced using the anti-ICOS Fab from XENP16435 and the anti-CTLA-4 scFv from ABD described in Example 2B. An exemplary anti-ICOS × anti-CTLA-4 bottle opener array is shown in FIG. Bispecific mAbs were produced as generally described in Example 5I (c).

2.6B: anti-LAG-3 × anti-ICOS
Anti-LAG-3 × anti-ICOS bispecific antibodies were produced using the anti-LAG-3 Fab from ABD and the anti-ICOS Fab from XENP16435 as described in Example 2C. The sequence of an exemplary anti-LAG-3x anti-ICOS bispecific antibody is depicted in FIG. Bispecific mAbs were prepared as generally described in Example 5I (c).

3.6C: anti-TIM-3 × anti-ICOS
Anti-TIM-3 × anti-ICOS bispecific antibodies were produced using anti-TIM-3 Fab from ABD and anti-ICOS Fab from XENP16435 as described in Example 2D. The sequence of an exemplary anti-TIM-3 × anti-ICOS bispecific antibody is shown in FIG. Bispecific mAbs were produced as generally described in Example 5I (c).

4.6D: Anti-PD-L1 × anti-ICOS
An anti-PD-L1 × anti-ICOS bottle opener was produced using the ABD-derived anti-PD-L1 Fab described in Example 2E and the anti-ICOS scFv described in Example 4B. The sequence of an exemplary anti-PD-L1 × anti-ICOS bispecific antibody is shown in FIG. A bottle opener was produced as generally described in Example 5A.

5.6E: Additional costim x checkpoint blocking bispecific antibody functions to enhance cytokine secretion Human PBMC were stimulated with 200 ng / mL SEB for 2 days. Cells were washed twice and restimulated with 100 ng / mL SEB in combination with the indicated test samples at the indicated concentrations (as described above). After 24 hours of treatment, the supernatant was assayed for IL-2. Data are shown in FIG. 67 and show that each additional costim × checkpoint blocking bispecific antibody enhances IL-2 secretion compared to the bivalent anti-RSV (XENP 15074) control. In particular, the majority of these alternative checkpoint blocking antibodies (eg, but not all TIM-3 blocking) enhance IL-2 secretion compared to a bivalent anti-PD-1 antibody (XENP16432). .

G. Example 7: Monovalent linking of ICOS is superior to bivalent crosslinking In Example 5A (a), surprisingly, anti-ICOS × anti-PD-1 antibody is required for stimulation of cytokine production It was found that it acts to increase cytokine secretion despite the monovalent linkage of ICOS opposite to that of costimulatory receptors such as ICOS, which is believed to be. In Example 5H (a), it was surprisingly found that the anti-ICOS × anti-PD-1 antibody enhances AKT phosphorylation (sign of ICOS agonism) over the bivalent anti-ICOS mAb. In this chapter, we further investigate this trend that ICOS monovalent consolidation appears to be superior to ICOS bivalent consolidation.

1.7A: Production of 1-arm anti-ICOS Fab-Fc antibody The amino acid sequence of an exemplary 1-arm anti-ICOS Fab-Fc antibody is listed in FIG. DNA encoding the three strands required for antibody expression was generated by gene synthesis and subcloned into the expression vector pTT5 using standard molecular biology techniques. HEK293E cells were DNA transfected for expression and the resulting protein was purified using standard techniques.

  The resulting 1-arm anti-ICOS Fab-Fc antibody was analyzed for affinity for human ICOS using Octet. An anti-mouse Fc (AMC) biosensor was used to capture the mouse IgG2a Fc fusion of ICOS and soak in multiple concentrations of test samples. The resulting equilibrium dissociation constant (KD), association rate (ka), and dissociation rate (kd) are presented in FIG. Ligation affinity and dynamic rate constants were obtained by analyzing processed data using a 1: 1 ligation model using ForteBio Octet Data Analysis software (ForteBio).

2.7B: Monovalent 1-arm anti-ICOS Fab-Fc antibody promotes greater AKT activation in PBMC than bivalent anti-ICOS antibody PBMC were stimulated with 100 ng / mL SEB for 2 days. After stimulation, CD3 + T cells were isolated by negative selection using EasySep ™ human T cell enrichment kit (STEMCELL Technologies, Vancouver, Canada) and then shown in combination with plate-linked anti-CD3 antibody (OKT3, 500 ng / mL) The test sample was treated. Cells were lysed 30 minutes after treatment and assayed for total AKT and phosphorylated AKT (Ser473) by multiple phosphoprotein assay on MULTI-SPOT 384 well spot plates (Meso Scale Discovery, Rockville, Md.). Data is shown in FIG. 70 as a percentage of AKT phosphorylated after treatment.

  Consistent with Example X, the anti-ICOS × anti-PD-1 bispecific antibody promoted significantly higher AKT activation than the bivalent anti-ICOS antibody. Surprisingly, the enhanced monovalent 1-arm anti-ICOS Fab-Fc antibody was also significantly higher than the bivalent anti-ICOS antibody and promoted AKT activation to a level comparable to bispecific antibodies. .

3.7C: ICOS monovalent agonism coordinates with multiple anti-ICOS ABD CD3 + T cells were isolated by negative selection using EasySep ™ human T cell enrichment kit (STEMCELL Technologies, Vancouver, Canada) and then plate ligated Treated with the indicated test samples in combination with anti-CD3 antibody (OKT3, 500 ng / mL). Cells were lysed 30 minutes after treatment and assayed for total AKT and phosphorylated AKT (Ser473) by multiple phosphoprotein assay on MULTI-SPOT 384 well spot plates (Meso Scale Discovery, Rockville, Md.). The data is shown in FIG. 74 and indicates that monovalent anti-ICOS Fab-Fc antibodies containing various anti-ICOS ABDs were able to induce AKT phosphorylation.

Claims (25)

  A bispecific antibody that is monovalently linked to a human costimulatory receptor and monovalently linked to a human checkpoint receptor for use in activating T cells for the treatment of cancer.   The bispecific antibody of claim 1, wherein the costimulatory receptor is selected from the group consisting of ICOS, GITR, OX40, and 4-1BB.   The bispecific antibody of claim 1 or 2, wherein the checkpoint receptor is selected from the group consisting of PD-1, PD-L1, CTLA-4, LAG-3, and TIM-3.   4. The bispecific antibody according to claim 1, 2 or 3, wherein the costimulatory receptor is human ICOS.   5. The bispecific antibody of claim 4, wherein the antibody is linked to ICOS and PD-1.   5. The bispecific antibody of claim 4, wherein the antibody is linked to ICOS and PD-L1.   5. The bispecific antibody of claim 4, wherein the antibody is linked to ICOS and CTLA-4.   5. The bispecific antibody of claim 4, wherein the antibody is linked to ICOS and LAG-3.   5. The bispecific antibody of claim 4, wherein the antibody is linked to ICOS and TIM-3.   10. The bispecific antibody according to any of claims 1-9, wherein the antibody has the format of Figure 2A.   The bispecific antibody of any of claims 1 to 10, wherein the antibody has the format of Figure 2E.   12. The bispecific antibody according to any of claims 1 to 11, wherein the antibody has the format of Figure 2F.   The bispecific antibody according to any of claims 1 to 12, wherein the antibody has the format of Fig. 2G. The antibody is
a) a first sequence having SEQ ID NO: 26642,
6. The bispecific antibody of claim 5, comprising b) a second sequence having SEQ ID NO: 26657, and c) a third sequence having SEQ ID NO: 26647. A heterodimeric antibody comprising:
a) a first heavy chain comprising a first antigen-binding domain comprising a first Fc domain, an optional domain linker, and an scFv linking to the first antigen;
b) a second heavy chain comprising a second Fc domain, a hinge domain, a heavy chain constant domain comprising a CH1 domain, and a heavy chain comprising a variable heavy chain domain;
c) a light chain comprising a variable light chain domain and a light chain constant domain, and
The variable heavy chain domain and the variable light chain domain form a second antigen linking domain linked to a second antigen, and one of the first and second antigen linking domains is linked to human ICOS. The other is a heterodimeric antibody linked to human PD-1. A heterodimeric antibody comprising:
a) the first heavy chain,
i) a first variant Fc domain, and ii) a single chain Fv region (scFv) linked to a first antigen, wherein the scFv region comprises a first variable heavy chain, a variable light chain, and a charged scFv linker A first heavy chain, wherein the charged scFv linker covalently links the first variable heavy chain and the variable light chain;
b) a second heavy chain comprising a VH-CH1-hinge-CH2-CH3 monomer, where VH is a second variable heavy chain and CH2-CH3 is a second variant Fc domain; A second heavy chain;
c) a light chain, and
The second variant Fc domain comprises the amino acid substitution N208D / Q295E / N384D / Q418E / N241D;
The first and second variant Fc domains each comprise the amino acid substitution E233P / L234V / L235A / G236del / S267K;
The first mutant Fc domain comprises the amino acid substitution S364K / E357Q, the second mutant Fc domain comprises the amino acid substitution L368D / K370S, and one of the first and second antigen-binding domains is human ICOS Heterodimeric antibody with the other linked to human PD-1 and numbering according to the EU index similar to Kabat.   The heterodimeric antibody according to any of claims 1 to 16, wherein the first and second variant Fc domains each comprise M428L / N434S. A heterodimeric antibody comprising:
a) the first heavy chain,
i) a first variant Fc domain, and ii) a single chain Fv region (scFv) linked to a first antigen, wherein the scFv region comprises a first variable heavy chain, a variable light chain, and a charged scFv linker A first heavy chain, wherein the charged scFv linker covalently links the first variable heavy chain and the variable light chain;
b) a second heavy chain comprising a VH-CH1-hinge-CH2-CH3 monomer, where VH is a second variable heavy chain and CH2-CH3 is a second variant Fc domain; A second heavy chain;
c) a light chain, and
The first and second variant Fc domains are L368D / K370S: S364K / E357Q, L368D / K370S: S364K, L368E / K370S: S364K, T411E / K360E / Q362E: D401K, and T366S / L368T / 36B: A set of heterodimerization mutations selected from the group consisting of: one of the first and second antigen-binding domains linked to human ICOS, the other linked to human PD-1, and numbering as Kabat and Heterodimeric antibody according to similar EU index. A nucleic acid composition comprising:
a) a first nucleic acid encoding the first heavy chain according to any of claims 1 to 18, and
b) a second nucleic acid encoding the second heavy chain according to any one of claims 1 to 18, and
c) a third nucleic acid encoding each of the light chains according to any one of claims 1 to 18, and a nucleic acid composition. An expression vector composition comprising:
a) a first expression vector comprising the first nucleic acid of claim 19;
b) a second expression vector comprising the second nucleic acid of claim 19;
c) a third expression vector comprising the third nucleic acid of claim 19, and an expression vector composition.   A host cell comprising the expression vector composition of claim 20.   A method for producing a heterodimeric antibody, comprising culturing the host cell of claim 21 under conditions under which the heterodimeric antibody is expressed, and recovering the antibody. Method.   A method of activating cytotoxic T cells (CTL) in a patient, comprising administering to said patient a heterodimeric antibody according to any of claims 1-18.   A method for increasing IFNγ levels in a patient comprising administering to said patient a heterodimeric antibody according to any of claims 1-18. A method of increasing IL-2 levels in a patient comprising administering to said patient a heterodimeric antibody according to any of claims 1-18.